sys/dev/netif/igb/if_igb.c

   1 /*
   2  * Copyright (c) 2001-2011, Intel Corporation
   3  * All rights reserved.
   4  *
   5  * Redistribution and use in source and binary forms, with or without
   6  * modification, are permitted provided that the following conditions are met:
   7  *
   8  *  1. Redistributions of source code must retain the above copyright notice,
   9  *     this list of conditions and the following disclaimer.
  10  *
  11  *  2. Redistributions in binary form must reproduce the above copyright
  12  *     notice, this list of conditions and the following disclaimer in the
  13  *     documentation and/or other materials provided with the distribution.
  14  *
  15  *  3. Neither the name of the Intel Corporation nor the names of its
  16  *     contributors may be used to endorse or promote products derived from
  17  *     this software without specific prior written permission.
  18  *
  19  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
  20  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  22  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
  23  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
  24  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
  25  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
  26  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
  27  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
  28  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
  29  * POSSIBILITY OF SUCH DAMAGE.
  30  */
  31
  32 #include "opt_ifpoll.h"
  33 #include "opt_igb.h"
  34
  35 #include <sys/param.h>
  36 #include <sys/bus.h>
  37 #include <sys/endian.h>
  38 #include <sys/interrupt.h>
  39 #include <sys/kernel.h>
  40 #include <sys/malloc.h>
  41 #include <sys/mbuf.h>
  42 #include <sys/proc.h>
  43 #include <sys/rman.h>
  44 #include <sys/serialize.h>
  45 #include <sys/serialize2.h>
  46 #include <sys/socket.h>
  47 #include <sys/sockio.h>
  48 #include <sys/sysctl.h>
  49 #include <sys/systm.h>
  50
  51 #include <net/bpf.h>
  52 #include <net/ethernet.h>
  53 #include <net/if.h>
  54 #include <net/if_arp.h>
  55 #include <net/if_dl.h>
  56 #include <net/if_media.h>
  57 #include <net/ifq_var.h>
  58 #include <net/toeplitz.h>
  59 #include <net/toeplitz2.h>
  60 #include <net/vlan/if_vlan_var.h>
  61 #include <net/vlan/if_vlan_ether.h>
  62 #include <net/if_poll.h>
  63
  64 #include <netinet/in_systm.h>
  65 #include <netinet/in.h>
  66 #include <netinet/ip.h>
  67
  68 #include <bus/pci/pcivar.h>
  69 #include <bus/pci/pcireg.h>
  70
  71 #include <dev/netif/ig_hal/e1000_api.h>
  72 #include <dev/netif/ig_hal/e1000_82575.h>
  73 #include <dev/netif/igb/if_igb.h>
  74
  75 #ifdef IGB_RSS_DEBUG
  76 #define IGB_RSS_DPRINTF(sc, lvl, fmt, ...) \
  77 do { \
  78         if (sc->rss_debug >= lvl) \
  79                 if_printf(&sc->arpcom.ac_if, fmt, __VA_ARGS__); \
  80 } while (0)
  81 #else   /* !IGB_RSS_DEBUG */
  82 #define IGB_RSS_DPRINTF(sc, lvl, fmt, ...)      ((void)0)
  83 #endif  /* IGB_RSS_DEBUG */
  84
  85 #define IGB_NAME        "Intel(R) PRO/1000 "
  86 #define IGB_DEVICE(id)  \
  87         { IGB_VENDOR_ID, E1000_DEV_ID_##id, IGB_NAME #id }
  88 #define IGB_DEVICE_NULL { 0, 0, NULL }
  89
  90 static struct igb_device {
  91         uint16_t        vid;
  92         uint16_t        did;
  93         const char      *desc;
  94 } igb_devices[] = {
  95         IGB_DEVICE(82575EB_COPPER),
  96         IGB_DEVICE(82575EB_FIBER_SERDES),
  97         IGB_DEVICE(82575GB_QUAD_COPPER),
  98         IGB_DEVICE(82576),
  99         IGB_DEVICE(82576_NS),
 100         IGB_DEVICE(82576_NS_SERDES),
 101         IGB_DEVICE(82576_FIBER),
 102         IGB_DEVICE(82576_SERDES),
 103         IGB_DEVICE(82576_SERDES_QUAD),
 104         IGB_DEVICE(82576_QUAD_COPPER),
 105         IGB_DEVICE(82576_QUAD_COPPER_ET2),
 106         IGB_DEVICE(82576_VF),
 107         IGB_DEVICE(82580_COPPER),
 108         IGB_DEVICE(82580_FIBER),
 109         IGB_DEVICE(82580_SERDES),
 110         IGB_DEVICE(82580_SGMII),
 111         IGB_DEVICE(82580_COPPER_DUAL),
 112         IGB_DEVICE(82580_QUAD_FIBER),
 113         IGB_DEVICE(DH89XXCC_SERDES),
 114         IGB_DEVICE(DH89XXCC_SGMII),
 115         IGB_DEVICE(DH89XXCC_SFP),
 116         IGB_DEVICE(DH89XXCC_BACKPLANE),
 117         IGB_DEVICE(I350_COPPER),
 118         IGB_DEVICE(I350_FIBER),
 119         IGB_DEVICE(I350_SERDES),
 120         IGB_DEVICE(I350_SGMII),
 121         IGB_DEVICE(I350_VF),
 122         IGB_DEVICE(I210_COPPER),
 123         IGB_DEVICE(I210_COPPER_IT),
 124         IGB_DEVICE(I210_COPPER_OEM1),
 125         IGB_DEVICE(I210_COPPER_FLASHLESS),
 126         IGB_DEVICE(I210_SERDES_FLASHLESS),
 127         IGB_DEVICE(I210_FIBER),
 128         IGB_DEVICE(I210_SERDES),
 129         IGB_DEVICE(I210_SGMII),
 130         IGB_DEVICE(I211_COPPER),
 131         IGB_DEVICE(I354_BACKPLANE_1GBPS),
 132         IGB_DEVICE(I354_SGMII),
 133
 134         /* required last entry */
 135         IGB_DEVICE_NULL
 136 };
 137
 138 static int      igb_probe(device_t);
 139 static int      igb_attach(device_t);
 140 static int      igb_detach(device_t);
 141 static int      igb_shutdown(device_t);
 142 static int      igb_suspend(device_t);
 143 static int      igb_resume(device_t);
 144
 145 static boolean_t igb_is_valid_ether_addr(const uint8_t *);
 146 static void     igb_setup_ifp(struct igb_softc *);
 147 static boolean_t igb_txcsum_ctx(struct igb_tx_ring *, struct mbuf *);
 148 static int      igb_tso_pullup(struct igb_tx_ring *, struct mbuf **);
 149 static void     igb_tso_ctx(struct igb_tx_ring *, struct mbuf *, uint32_t *);
 150 static void     igb_add_sysctl(struct igb_softc *);
 151 static int      igb_sysctl_intr_rate(SYSCTL_HANDLER_ARGS);
 152 static int      igb_sysctl_msix_rate(SYSCTL_HANDLER_ARGS);
 153 static int      igb_sysctl_tx_intr_nsegs(SYSCTL_HANDLER_ARGS);
 154 static int      igb_sysctl_tx_wreg_nsegs(SYSCTL_HANDLER_ARGS);
 155 static int      igb_sysctl_rx_wreg_nsegs(SYSCTL_HANDLER_ARGS);
 156 static void     igb_set_ring_inuse(struct igb_softc *, boolean_t);
 157 static int      igb_get_rxring_inuse(const struct igb_softc *, boolean_t);
 158 static int      igb_get_txring_inuse(const struct igb_softc *, boolean_t);
 159 static void     igb_set_timer_cpuid(struct igb_softc *, boolean_t);
 160 #ifdef IFPOLL_ENABLE
 161 static int      igb_sysctl_npoll_rxoff(SYSCTL_HANDLER_ARGS);
 162 static int      igb_sysctl_npoll_txoff(SYSCTL_HANDLER_ARGS);
 163 #endif
 164
 165 static void     igb_vf_init_stats(struct igb_softc *);
 166 static void     igb_reset(struct igb_softc *);
 167 static void     igb_update_stats_counters(struct igb_softc *);
 168 static void     igb_update_vf_stats_counters(struct igb_softc *);
 169 static void     igb_update_link_status(struct igb_softc *);
 170 static void     igb_init_tx_unit(struct igb_softc *);
 171 static void     igb_init_rx_unit(struct igb_softc *);
 172
 173 static void     igb_set_vlan(struct igb_softc *);
 174 static void     igb_set_multi(struct igb_softc *);
 175 static void     igb_set_promisc(struct igb_softc *);
 176 static void     igb_disable_promisc(struct igb_softc *);
 177
 178 static int      igb_alloc_rings(struct igb_softc *);
 179 static void     igb_free_rings(struct igb_softc *);
 180 static int      igb_create_tx_ring(struct igb_tx_ring *);
 181 static int      igb_create_rx_ring(struct igb_rx_ring *);
 182 static void     igb_free_tx_ring(struct igb_tx_ring *);
 183 static void     igb_free_rx_ring(struct igb_rx_ring *);
 184 static void     igb_destroy_tx_ring(struct igb_tx_ring *, int);
 185 static void     igb_destroy_rx_ring(struct igb_rx_ring *, int);
 186 static void     igb_init_tx_ring(struct igb_tx_ring *);
 187 static int      igb_init_rx_ring(struct igb_rx_ring *);
 188 static int      igb_newbuf(struct igb_rx_ring *, int, boolean_t);
 189 static int      igb_encap(struct igb_tx_ring *, struct mbuf **, int *, int *);
 190 static void     igb_rx_refresh(struct igb_rx_ring *, int);
 191 static void     igb_setup_serializer(struct igb_softc *);
 192
 193 static void     igb_stop(struct igb_softc *);
 194 static void     igb_init(void *);
 195 static int      igb_ioctl(struct ifnet *, u_long, caddr_t, struct ucred *);
 196 static void     igb_media_status(struct ifnet *, struct ifmediareq *);
 197 static int      igb_media_change(struct ifnet *);
 198 static void     igb_timer(void *);
 199 static void     igb_watchdog(struct ifaltq_subque *);
 200 static void     igb_start(struct ifnet *, struct ifaltq_subque *);
 201 #ifdef IFPOLL_ENABLE
 202 static void     igb_npoll(struct ifnet *, struct ifpoll_info *);
 203 static void     igb_npoll_rx(struct ifnet *, void *, int);
 204 static void     igb_npoll_tx(struct ifnet *, void *, int);
 205 static void     igb_npoll_status(struct ifnet *);
 206 #endif
 207 static void     igb_serialize(struct ifnet *, enum ifnet_serialize);
 208 static void     igb_deserialize(struct ifnet *, enum ifnet_serialize);
 209 static int      igb_tryserialize(struct ifnet *, enum ifnet_serialize);
 210 #ifdef INVARIANTS
 211 static void     igb_serialize_assert(struct ifnet *, enum ifnet_serialize,
 212                     boolean_t);
 213 #endif
 214
 215 static void     igb_intr(void *);
 216 static void     igb_intr_shared(void *);
 217 static void     igb_rxeof(struct igb_rx_ring *, int);
 218 static void     igb_txeof(struct igb_tx_ring *);
 219 static void     igb_set_eitr(struct igb_softc *, int, int);
 220 static void     igb_enable_intr(struct igb_softc *);
 221 static void     igb_disable_intr(struct igb_softc *);
 222 static void     igb_init_unshared_intr(struct igb_softc *);
 223 static void     igb_init_intr(struct igb_softc *);
 224 static int      igb_setup_intr(struct igb_softc *);
 225 static void     igb_set_txintr_mask(struct igb_tx_ring *, int *, int);
 226 static void     igb_set_rxintr_mask(struct igb_rx_ring *, int *, int);
 227 static void     igb_set_intr_mask(struct igb_softc *);
 228 static int      igb_alloc_intr(struct igb_softc *);
 229 static void     igb_free_intr(struct igb_softc *);
 230 static void     igb_teardown_intr(struct igb_softc *);
 231 static void     igb_msix_try_alloc(struct igb_softc *);
 232 static void     igb_msix_rx_conf(struct igb_softc *, int, int *, int);
 233 static void     igb_msix_tx_conf(struct igb_softc *, int, int *, int);
 234 static void     igb_msix_free(struct igb_softc *, boolean_t);
 235 static int      igb_msix_setup(struct igb_softc *);
 236 static void     igb_msix_teardown(struct igb_softc *, int);
 237 static void     igb_msix_rx(void *);
 238 static void     igb_msix_tx(void *);
 239 static void     igb_msix_status(void *);
 240 static void     igb_msix_rxtx(void *);
 241
 242 /* Management and WOL Support */
 243 static void     igb_get_mgmt(struct igb_softc *);
 244 static void     igb_rel_mgmt(struct igb_softc *);
 245 static void     igb_get_hw_control(struct igb_softc *);
 246 static void     igb_rel_hw_control(struct igb_softc *);
 247 static void     igb_enable_wol(device_t);
 248
 249 static device_method_t igb_methods[] = {
 250         /* Device interface */
 251         DEVMETHOD(device_probe,         igb_probe),
 252         DEVMETHOD(device_attach,        igb_attach),
 253         DEVMETHOD(device_detach,        igb_detach),
 254         DEVMETHOD(device_shutdown,      igb_shutdown),
 255         DEVMETHOD(device_suspend,       igb_suspend),
 256         DEVMETHOD(device_resume,        igb_resume),
 257         DEVMETHOD_END
 258 };
 259
 260 static driver_t igb_driver = {
 261         "igb",
 262         igb_methods,
 263         sizeof(struct igb_softc),
 264 };
 265
 266 static devclass_t igb_devclass;
 267
 268 DECLARE_DUMMY_MODULE(if_igb);
 269 MODULE_DEPEND(igb, ig_hal, 1, 1, 1);
 270 DRIVER_MODULE(if_igb, pci, igb_driver, igb_devclass, NULL, NULL);
 271
 272 static int      igb_rxd = IGB_DEFAULT_RXD;
 273 static int      igb_txd = IGB_DEFAULT_TXD;
 274 static int      igb_rxr = 0;
 275 static int      igb_txr = 0;
 276 static int      igb_msi_enable = 1;
 277 static int      igb_msix_enable = 1;
 278 static int      igb_eee_disabled = 1;   /* Energy Efficient Ethernet */
 279 static int      igb_fc_setting = e1000_fc_full;
 280
 281 /*
 282  * DMA Coalescing, only for i350 - default to off,
 283  * this feature is for power savings
 284  */
 285 static int      igb_dma_coalesce = 0;
 286
 287 TUNABLE_INT("hw.igb.rxd", &igb_rxd);
 288 TUNABLE_INT("hw.igb.txd", &igb_txd);
 289 TUNABLE_INT("hw.igb.rxr", &igb_rxr);
 290 TUNABLE_INT("hw.igb.txr", &igb_txr);
 291 TUNABLE_INT("hw.igb.msi.enable", &igb_msi_enable);
 292 TUNABLE_INT("hw.igb.msix.enable", &igb_msix_enable);
 293 TUNABLE_INT("hw.igb.fc_setting", &igb_fc_setting);
 294
 295 /* i350 specific */
 296 TUNABLE_INT("hw.igb.eee_disabled", &igb_eee_disabled);
 297 TUNABLE_INT("hw.igb.dma_coalesce", &igb_dma_coalesce);
 298
 299 static __inline void
 300 igb_rxcsum(uint32_t staterr, struct mbuf *mp)
 301 {
 302         /* Ignore Checksum bit is set */
 303         if (staterr & E1000_RXD_STAT_IXSM)
 304                 return;
 305
 306         if ((staterr & (E1000_RXD_STAT_IPCS | E1000_RXDEXT_STATERR_IPE)) ==
 307             E1000_RXD_STAT_IPCS)
 308                 mp->m_pkthdr.csum_flags |= CSUM_IP_CHECKED | CSUM_IP_VALID;
 309
 310         if (staterr & (E1000_RXD_STAT_TCPCS | E1000_RXD_STAT_UDPCS)) {
 311                 if ((staterr & E1000_RXDEXT_STATERR_TCPE) == 0) {
 312                         mp->m_pkthdr.csum_flags |= CSUM_DATA_VALID |
 313                             CSUM_PSEUDO_HDR | CSUM_FRAG_NOT_CHECKED;
 314                         mp->m_pkthdr.csum_data = htons(0xffff);
 315                 }
 316         }
 317 }
 318
 319 static __inline struct pktinfo *
 320 igb_rssinfo(struct mbuf *m, struct pktinfo *pi,
 321     uint32_t hash, uint32_t hashtype, uint32_t staterr)
 322 {
 323         switch (hashtype) {
 324         case E1000_RXDADV_RSSTYPE_IPV4_TCP:
 325                 pi->pi_netisr = NETISR_IP;
 326                 pi->pi_flags = 0;
 327                 pi->pi_l3proto = IPPROTO_TCP;
 328                 break;
 329
 330         case E1000_RXDADV_RSSTYPE_IPV4:
 331                 if (staterr & E1000_RXD_STAT_IXSM)
 332                         return NULL;
 333
 334                 if ((staterr &
 335                      (E1000_RXD_STAT_TCPCS | E1000_RXDEXT_STATERR_TCPE)) ==
 336                     E1000_RXD_STAT_TCPCS) {
 337                         pi->pi_netisr = NETISR_IP;
 338                         pi->pi_flags = 0;
 339                         pi->pi_l3proto = IPPROTO_UDP;
 340                         break;
 341                 }
 342                 /* FALL THROUGH */
 343         default:
 344                 return NULL;
 345         }
 346
 347         m->m_flags |= M_HASH;
 348         m->m_pkthdr.hash = toeplitz_hash(hash);
 349         return pi;
 350 }
 351
 352 static int
 353 igb_probe(device_t dev)
 354 {
 355         const struct igb_device *d;
 356         uint16_t vid, did;
 357
 358         vid = pci_get_vendor(dev);
 359         did = pci_get_device(dev);
 360
 361         for (d = igb_devices; d->desc != NULL; ++d) {
 362                 if (vid == d->vid && did == d->did) {
 363                         device_set_desc(dev, d->desc);
 364                         return 0;
 365                 }
 366         }
 367         return ENXIO;
 368 }
 369
 370 static int
 371 igb_attach(device_t dev)
 372 {
 373         struct igb_softc *sc = device_get_softc(dev);
 374         uint16_t eeprom_data;
 375         int error = 0, ring_max;
 376 #ifdef IFPOLL_ENABLE
 377         int offset, offset_def;
 378 #endif
 379
 380 #ifdef notyet
 381         /* SYSCTL stuff */
 382         SYSCTL_ADD_PROC(device_get_sysctl_ctx(dev),
 383             SYSCTL_CHILDREN(device_get_sysctl_tree(dev)),
 384             OID_AUTO, "nvm", CTLTYPE_INT|CTLFLAG_RW, adapter, 0,
 385             igb_sysctl_nvm_info, "I", "NVM Information");
 386         SYSCTL_ADD_PROC(device_get_sysctl_ctx(dev),
 387             SYSCTL_CHILDREN(device_get_sysctl_tree(dev)),
 388             OID_AUTO, "flow_control", CTLTYPE_INT|CTLFLAG_RW,
 389             adapter, 0, igb_set_flowcntl, "I", "Flow Control");
 390 #endif
 391
 392         callout_init_mp(&sc->timer);
 393         lwkt_serialize_init(&sc->main_serialize);
 394
 395         if_initname(&sc->arpcom.ac_if, device_get_name(dev),
 396             device_get_unit(dev));
 397         sc->dev = sc->osdep.dev = dev;
 398
 399         /*
 400          * Determine hardware and mac type
 401          */
 402         sc->hw.vendor_id = pci_get_vendor(dev);
 403         sc->hw.device_id = pci_get_device(dev);
 404         sc->hw.revision_id = pci_read_config(dev, PCIR_REVID, 1);
 405         sc->hw.subsystem_vendor_id = pci_read_config(dev, PCIR_SUBVEND_0, 2);
 406         sc->hw.subsystem_device_id = pci_read_config(dev, PCIR_SUBDEV_0, 2);
 407
 408         if (e1000_set_mac_type(&sc->hw))
 409                 return ENXIO;
 410
 411         /* Are we a VF device? */
 412         if (sc->hw.mac.type == e1000_vfadapt ||
 413             sc->hw.mac.type == e1000_vfadapt_i350)
 414                 sc->vf_ifp = 1;
 415         else
 416                 sc->vf_ifp = 0;
 417
 418         /*
 419          * Configure total supported RX/TX ring count
 420          */
 421         switch (sc->hw.mac.type) {
 422         case e1000_82575:
 423                 ring_max = IGB_MAX_RING_82575;
 424                 break;
 425
 426         case e1000_82576:
 427                 ring_max = IGB_MAX_RING_82576;
 428                 break;
 429
 430         case e1000_82580:
 431                 ring_max = IGB_MAX_RING_82580;
 432                 break;
 433
 434         case e1000_i350:
 435                 ring_max = IGB_MAX_RING_I350;
 436                 break;
 437
 438         case e1000_i354:
 439                 ring_max = IGB_MAX_RING_I354;
 440                 break;
 441
 442         case e1000_i210:
 443                 ring_max = IGB_MAX_RING_I210;
 444                 break;
 445
 446         case e1000_i211:
 447                 ring_max = IGB_MAX_RING_I211;
 448                 break;
 449
 450         default:
 451                 ring_max = IGB_MIN_RING;
 452                 break;
 453         }
 454
 455         sc->rx_ring_cnt = device_getenv_int(dev, "rxr", igb_rxr);
 456         sc->rx_ring_cnt = if_ring_count2(sc->rx_ring_cnt, ring_max);
 457 #ifdef IGB_RSS_DEBUG
 458         sc->rx_ring_cnt = device_getenv_int(dev, "rxr_debug", sc->rx_ring_cnt);
 459 #endif
 460         sc->rx_ring_inuse = sc->rx_ring_cnt;
 461
 462         sc->tx_ring_cnt = device_getenv_int(dev, "txr", igb_txr);
 463         sc->tx_ring_cnt = if_ring_count2(sc->tx_ring_cnt, ring_max);
 464 #ifdef IGB_TSS_DEBUG
 465         sc->tx_ring_cnt = device_getenv_int(dev, "txr_debug", sc->tx_ring_cnt);
 466 #endif
 467         sc->tx_ring_inuse = sc->tx_ring_cnt;
 468
 469         /* Enable bus mastering */
 470         pci_enable_busmaster(dev);
 471
 472         /*
 473          * Allocate IO memory
 474          */
 475         sc->mem_rid = PCIR_BAR(0);
 476         sc->mem_res = bus_alloc_resource_any(dev, SYS_RES_MEMORY, &sc->mem_rid,
 477             RF_ACTIVE);
 478         if (sc->mem_res == NULL) {
 479                 device_printf(dev, "Unable to allocate bus resource: memory\n");
 480                 error = ENXIO;
 481                 goto failed;
 482         }
 483         sc->osdep.mem_bus_space_tag = rman_get_bustag(sc->mem_res);
 484         sc->osdep.mem_bus_space_handle = rman_get_bushandle(sc->mem_res);
 485
 486         sc->hw.hw_addr = (uint8_t *)&sc->osdep.mem_bus_space_handle;
 487
 488         /* Save PCI command register for Shared Code */
 489         sc->hw.bus.pci_cmd_word = pci_read_config(dev, PCIR_COMMAND, 2);
 490         sc->hw.back = &sc->osdep;
 491
 492         /* Do Shared Code initialization */
 493         if (e1000_setup_init_funcs(&sc->hw, TRUE)) {
 494                 device_printf(dev, "Setup of Shared code failed\n");
 495                 error = ENXIO;
 496                 goto failed;
 497         }
 498
 499         e1000_get_bus_info(&sc->hw);
 500
 501         sc->hw.mac.autoneg = DO_AUTO_NEG;
 502         sc->hw.phy.autoneg_wait_to_complete = FALSE;
 503         sc->hw.phy.autoneg_advertised = AUTONEG_ADV_DEFAULT;
 504
 505         /* Copper options */
 506         if (sc->hw.phy.media_type == e1000_media_type_copper) {
 507                 sc->hw.phy.mdix = AUTO_ALL_MODES;
 508                 sc->hw.phy.disable_polarity_correction = FALSE;
 509                 sc->hw.phy.ms_type = IGB_MASTER_SLAVE;
 510         }
 511
 512         /* Set the frame limits assuming  standard ethernet sized frames. */
 513         sc->max_frame_size = ETHERMTU + ETHER_HDR_LEN + ETHER_CRC_LEN;
 514
 515         /* Allocate RX/TX rings */
 516         error = igb_alloc_rings(sc);
 517         if (error)
 518                 goto failed;
 519
 520 #ifdef IFPOLL_ENABLE
 521         /*
 522          * NPOLLING RX CPU offset
 523          */
 524         if (sc->rx_ring_cnt == ncpus2) {
 525                 offset = 0;
 526         } else {
 527                 offset_def = (sc->rx_ring_cnt * device_get_unit(dev)) % ncpus2;
 528                 offset = device_getenv_int(dev, "npoll.rxoff", offset_def);
 529                 if (offset >= ncpus2 ||
 530                     offset % sc->rx_ring_cnt != 0) {
 531                         device_printf(dev, "invalid npoll.rxoff %d, use %d\n",
 532                             offset, offset_def);
 533                         offset = offset_def;
 534                 }
 535         }
 536         sc->rx_npoll_off = offset;
 537
 538         /*
 539          * NPOLLING TX CPU offset
 540          */
 541         if (sc->tx_ring_cnt == ncpus2) {
 542                 offset = 0;
 543         } else {
 544                 offset_def = (sc->tx_ring_cnt * device_get_unit(dev)) % ncpus2;
 545                 offset = device_getenv_int(dev, "npoll.txoff", offset_def);
 546                 if (offset >= ncpus2 ||
 547                     offset % sc->tx_ring_cnt != 0) {
 548                         device_printf(dev, "invalid npoll.txoff %d, use %d\n",
 549                             offset, offset_def);
 550                         offset = offset_def;
 551                 }
 552         }
 553         sc->tx_npoll_off = offset;
 554 #endif
 555
 556         /* Allocate interrupt */
 557         error = igb_alloc_intr(sc);
 558         if (error)
 559                 goto failed;
 560
 561         /* Setup serializers */
 562         igb_setup_serializer(sc);
 563
 564         /* Allocate the appropriate stats memory */
 565         if (sc->vf_ifp) {
 566                 sc->stats = kmalloc(sizeof(struct e1000_vf_stats), M_DEVBUF,
 567                     M_WAITOK | M_ZERO);
 568                 igb_vf_init_stats(sc);
 569         } else {
 570                 sc->stats = kmalloc(sizeof(struct e1000_hw_stats), M_DEVBUF,
 571                     M_WAITOK | M_ZERO);
 572         }
 573
 574         /* Allocate multicast array memory. */
 575         sc->mta = kmalloc(ETHER_ADDR_LEN * MAX_NUM_MULTICAST_ADDRESSES,
 576             M_DEVBUF, M_WAITOK);
 577
 578         /* Some adapter-specific advanced features */
 579         if (sc->hw.mac.type >= e1000_i350) {
 580 #ifdef notyet
 581                 igb_set_sysctl_value(adapter, "dma_coalesce",
 582                     "configure dma coalesce",
 583                     &adapter->dma_coalesce, igb_dma_coalesce);
 584                 igb_set_sysctl_value(adapter, "eee_disabled",
 585                     "enable Energy Efficient Ethernet",
 586                     &adapter->hw.dev_spec._82575.eee_disable,
 587                     igb_eee_disabled);
 588 #else
 589                 sc->dma_coalesce = igb_dma_coalesce;
 590                 sc->hw.dev_spec._82575.eee_disable = igb_eee_disabled;
 591 #endif
 592                 if (sc->hw.phy.media_type == e1000_media_type_copper) {
 593                         if (sc->hw.mac.type == e1000_i354)
 594                                 e1000_set_eee_i354(&sc->hw);
 595                         else
 596                                 e1000_set_eee_i350(&sc->hw);
 597                 }
 598         }
 599
 600         /*
 601          * Start from a known state, this is important in reading the nvm and
 602          * mac from that.
 603          */
 604         e1000_reset_hw(&sc->hw);
 605
 606         /* Make sure we have a good EEPROM before we read from it */
 607         if (sc->hw.mac.type != e1000_i210 && sc->hw.mac.type != e1000_i211 &&
 608             e1000_validate_nvm_checksum(&sc->hw) < 0) {
 609                 /*
 610                  * Some PCI-E parts fail the first check due to
 611                  * the link being in sleep state, call it again,
 612                  * if it fails a second time its a real issue.
 613                  */
 614                 if (e1000_validate_nvm_checksum(&sc->hw) < 0) {
 615                         device_printf(dev,
 616                             "The EEPROM Checksum Is Not Valid\n");
 617                         error = EIO;
 618                         goto failed;
 619                 }
 620         }
 621
 622         /* Copy the permanent MAC address out of the EEPROM */
 623         if (e1000_read_mac_addr(&sc->hw) < 0) {
 624                 device_printf(dev, "EEPROM read error while reading MAC"
 625                     " address\n");
 626                 error = EIO;
 627                 goto failed;
 628         }
 629         if (!igb_is_valid_ether_addr(sc->hw.mac.addr)) {
 630                 device_printf(dev, "Invalid MAC address\n");
 631                 error = EIO;
 632                 goto failed;
 633         }
 634
 635         /* Setup OS specific network interface */
 636         igb_setup_ifp(sc);
 637
 638         /* Add sysctl tree, must after igb_setup_ifp() */
 639         igb_add_sysctl(sc);
 640
 641         /* Now get a good starting state */
 642         igb_reset(sc);
 643
 644         /* Initialize statistics */
 645         igb_update_stats_counters(sc);
 646
 647         sc->hw.mac.get_link_status = 1;
 648         igb_update_link_status(sc);
 649
 650         /* Indicate SOL/IDER usage */
 651         if (e1000_check_reset_block(&sc->hw)) {
 652                 device_printf(dev,
 653                     "PHY reset is blocked due to SOL/IDER session.\n");
 654         }
 655
 656         /* Determine if we have to control management hardware */
 657         if (e1000_enable_mng_pass_thru(&sc->hw))
 658                 sc->flags |= IGB_FLAG_HAS_MGMT;
 659
 660         /*
 661          * Setup Wake-on-Lan
 662          */
 663         /* APME bit in EEPROM is mapped to WUC.APME */
 664         eeprom_data = E1000_READ_REG(&sc->hw, E1000_WUC) & E1000_WUC_APME;
 665         if (eeprom_data)
 666                 sc->wol = E1000_WUFC_MAG;
 667         /* XXX disable WOL */
 668         sc->wol = 0;
 669
 670 #ifdef notyet
 671         /* Register for VLAN events */
 672         adapter->vlan_attach = EVENTHANDLER_REGISTER(vlan_config,
 673              igb_register_vlan, adapter, EVENTHANDLER_PRI_FIRST);
 674         adapter->vlan_detach = EVENTHANDLER_REGISTER(vlan_unconfig,
 675              igb_unregister_vlan, adapter, EVENTHANDLER_PRI_FIRST);
 676 #endif
 677
 678 #ifdef notyet
 679         igb_add_hw_stats(adapter);
 680 #endif
 681
 682         /*
 683          * Disable interrupt to prevent spurious interrupts (line based
 684          * interrupt, MSI or even MSI-X), which had been observed on
 685          * several types of LOMs, from being handled.
 686          */
 687         igb_disable_intr(sc);
 688
 689         error = igb_setup_intr(sc);
 690         if (error) {
 691                 ether_ifdetach(&sc->arpcom.ac_if);
 692                 goto failed;
 693         }
 694         return 0;
 695
 696 failed:
 697         igb_detach(dev);
 698         return error;
 699 }
 700
 701 static int
 702 igb_detach(device_t dev)
 703 {
 704         struct igb_softc *sc = device_get_softc(dev);
 705
 706         if (device_is_attached(dev)) {
 707                 struct ifnet *ifp = &sc->arpcom.ac_if;
 708
 709                 ifnet_serialize_all(ifp);
 710
 711                 igb_stop(sc);
 712
 713                 e1000_phy_hw_reset(&sc->hw);
 714
 715                 /* Give control back to firmware */
 716                 igb_rel_mgmt(sc);
 717                 igb_rel_hw_control(sc);
 718
 719                 if (sc->wol) {
 720                         E1000_WRITE_REG(&sc->hw, E1000_WUC, E1000_WUC_PME_EN);
 721                         E1000_WRITE_REG(&sc->hw, E1000_WUFC, sc->wol);
 722                         igb_enable_wol(dev);
 723                 }
 724
 725                 igb_teardown_intr(sc);
 726
 727                 ifnet_deserialize_all(ifp);
 728
 729                 ether_ifdetach(ifp);
 730         } else if (sc->mem_res != NULL) {
 731                 igb_rel_hw_control(sc);
 732         }
 733         bus_generic_detach(dev);
 734
 735         igb_free_intr(sc);
 736
 737         if (sc->msix_mem_res != NULL) {
 738                 bus_release_resource(dev, SYS_RES_MEMORY, sc->msix_mem_rid,
 739                     sc->msix_mem_res);
 740         }
 741         if (sc->mem_res != NULL) {
 742                 bus_release_resource(dev, SYS_RES_MEMORY, sc->mem_rid,
 743                     sc->mem_res);
 744         }
 745
 746         igb_free_rings(sc);
 747
 748         if (sc->mta != NULL)
 749                 kfree(sc->mta, M_DEVBUF);
 750         if (sc->stats != NULL)
 751                 kfree(sc->stats, M_DEVBUF);
 752         if (sc->serializes != NULL)
 753                 kfree(sc->serializes, M_DEVBUF);
 754
 755         return 0;
 756 }
 757
 758 static int
 759 igb_shutdown(device_t dev)
 760 {
 761         return igb_suspend(dev);
 762 }
 763
 764 static int
 765 igb_suspend(device_t dev)
 766 {
 767         struct igb_softc *sc = device_get_softc(dev);
 768         struct ifnet *ifp = &sc->arpcom.ac_if;
 769
 770         ifnet_serialize_all(ifp);
 771
 772         igb_stop(sc);
 773
 774         igb_rel_mgmt(sc);
 775         igb_rel_hw_control(sc);
 776
 777         if (sc->wol) {
 778                 E1000_WRITE_REG(&sc->hw, E1000_WUC, E1000_WUC_PME_EN);
 779                 E1000_WRITE_REG(&sc->hw, E1000_WUFC, sc->wol);
 780                 igb_enable_wol(dev);
 781         }
 782
 783         ifnet_deserialize_all(ifp);
 784
 785         return bus_generic_suspend(dev);
 786 }
 787
 788 static int
 789 igb_resume(device_t dev)
 790 {
 791         struct igb_softc *sc = device_get_softc(dev);
 792         struct ifnet *ifp = &sc->arpcom.ac_if;
 793         int i;
 794
 795         ifnet_serialize_all(ifp);
 796
 797         igb_init(sc);
 798         igb_get_mgmt(sc);
 799
 800         for (i = 0; i < sc->tx_ring_inuse; ++i)
 801                 ifsq_devstart_sched(sc->tx_rings[i].ifsq);
 802
 803         ifnet_deserialize_all(ifp);
 804
 805         return bus_generic_resume(dev);
 806 }
 807
 808 static int
 809 igb_ioctl(struct ifnet *ifp, u_long command, caddr_t data, struct ucred *cr)
 810 {
 811         struct igb_softc *sc = ifp->if_softc;
 812         struct ifreq *ifr = (struct ifreq *)data;
 813         int max_frame_size, mask, reinit;
 814         int error = 0;
 815
 816         ASSERT_IFNET_SERIALIZED_ALL(ifp);
 817
 818         switch (command) {
 819         case SIOCSIFMTU:
 820                 max_frame_size = 9234;
 821                 if (ifr->ifr_mtu > max_frame_size - ETHER_HDR_LEN -
 822                     ETHER_CRC_LEN) {
 823                         error = EINVAL;
 824                         break;
 825                 }
 826
 827                 ifp->if_mtu = ifr->ifr_mtu;
 828                 sc->max_frame_size = ifp->if_mtu + ETHER_HDR_LEN +
 829                     ETHER_CRC_LEN;
 830
 831                 if (ifp->if_flags & IFF_RUNNING)
 832                         igb_init(sc);
 833                 break;
 834
 835         case SIOCSIFFLAGS:
 836                 if (ifp->if_flags & IFF_UP) {
 837                         if (ifp->if_flags & IFF_RUNNING) {
 838                                 if ((ifp->if_flags ^ sc->if_flags) &
 839                                     (IFF_PROMISC | IFF_ALLMULTI)) {
 840                                         igb_disable_promisc(sc);
 841                                         igb_set_promisc(sc);
 842                                 }
 843                         } else {
 844                                 igb_init(sc);
 845                         }
 846                 } else if (ifp->if_flags & IFF_RUNNING) {
 847                         igb_stop(sc);
 848                 }
 849                 sc->if_flags = ifp->if_flags;
 850                 break;
 851
 852         case SIOCADDMULTI:
 853         case SIOCDELMULTI:
 854                 if (ifp->if_flags & IFF_RUNNING) {
 855                         igb_disable_intr(sc);
 856                         igb_set_multi(sc);
 857 #ifdef IFPOLL_ENABLE
 858                         if (!(ifp->if_flags & IFF_NPOLLING))
 859 #endif
 860                                 igb_enable_intr(sc);
 861                 }
 862                 break;
 863
 864         case SIOCSIFMEDIA:
 865                 /* Check SOL/IDER usage */
 866                 if (e1000_check_reset_block(&sc->hw)) {
 867                         if_printf(ifp, "Media change is "
 868                             "blocked due to SOL/IDER session.\n");
 869                         break;
 870                 }
 871                 /* FALL THROUGH */
 872
 873         case SIOCGIFMEDIA:
 874                 error = ifmedia_ioctl(ifp, ifr, &sc->media, command);
 875                 break;
 876
 877         case SIOCSIFCAP:
 878                 reinit = 0;
 879                 mask = ifr->ifr_reqcap ^ ifp->if_capenable;
 880                 if (mask & IFCAP_RXCSUM) {
 881                         ifp->if_capenable ^= IFCAP_RXCSUM;
 882                         reinit = 1;
 883                 }
 884                 if (mask & IFCAP_VLAN_HWTAGGING) {
 885                         ifp->if_capenable ^= IFCAP_VLAN_HWTAGGING;
 886                         reinit = 1;
 887                 }
 888                 if (mask & IFCAP_TXCSUM) {
 889                         ifp->if_capenable ^= IFCAP_TXCSUM;
 890                         if (ifp->if_capenable & IFCAP_TXCSUM)
 891                                 ifp->if_hwassist |= IGB_CSUM_FEATURES;
 892                         else
 893                                 ifp->if_hwassist &= ~IGB_CSUM_FEATURES;
 894                 }
 895                 if (mask & IFCAP_TSO) {
 896                         ifp->if_capenable ^= IFCAP_TSO;
 897                         if (ifp->if_capenable & IFCAP_TSO)
 898                                 ifp->if_hwassist |= CSUM_TSO;
 899                         else
 900                                 ifp->if_hwassist &= ~CSUM_TSO;
 901                 }
 902                 if (mask & IFCAP_RSS)
 903                         ifp->if_capenable ^= IFCAP_RSS;
 904                 if (reinit && (ifp->if_flags & IFF_RUNNING))
 905                         igb_init(sc);
 906                 break;
 907
 908         default:
 909                 error = ether_ioctl(ifp, command, data);
 910                 break;
 911         }
 912         return error;
 913 }
 914
 915 static void
 916 igb_init(void *xsc)
 917 {
 918         struct igb_softc *sc = xsc;
 919         struct ifnet *ifp = &sc->arpcom.ac_if;
 920         boolean_t polling;
 921         int i;
 922
 923         ASSERT_IFNET_SERIALIZED_ALL(ifp);
 924
 925         igb_stop(sc);
 926
 927         /* Get the latest mac address, User can use a LAA */
 928         bcopy(IF_LLADDR(ifp), sc->hw.mac.addr, ETHER_ADDR_LEN);
 929
 930         /* Put the address into the Receive Address Array */
 931         e1000_rar_set(&sc->hw, sc->hw.mac.addr, 0);
 932
 933         igb_reset(sc);
 934         igb_update_link_status(sc);
 935
 936         E1000_WRITE_REG(&sc->hw, E1000_VET, ETHERTYPE_VLAN);
 937
 938         /* Configure for OS presence */
 939         igb_get_mgmt(sc);
 940
 941         polling = FALSE;
 942 #ifdef IFPOLL_ENABLE
 943         if (ifp->if_flags & IFF_NPOLLING)
 944                 polling = TRUE;
 945 #endif
 946
 947         /* Configured used RX/TX rings */
 948         igb_set_ring_inuse(sc, polling);
 949         ifq_set_subq_mask(&ifp->if_snd, sc->tx_ring_inuse - 1);
 950
 951         /* Initialize interrupt */
 952         igb_init_intr(sc);
 953
 954         /* Prepare transmit descriptors and buffers */
 955         for (i = 0; i < sc->tx_ring_inuse; ++i)
 956                 igb_init_tx_ring(&sc->tx_rings[i]);
 957         igb_init_tx_unit(sc);
 958
 959         /* Setup Multicast table */
 960         igb_set_multi(sc);
 961
 962 #if 0
 963         /*
 964          * Figure out the desired mbuf pool
 965          * for doing jumbo/packetsplit
 966          */
 967         if (adapter->max_frame_size <= 2048)
 968                 adapter->rx_mbuf_sz = MCLBYTES;
 969         else if (adapter->max_frame_size <= 4096)
 970                 adapter->rx_mbuf_sz = MJUMPAGESIZE;
 971         else
 972                 adapter->rx_mbuf_sz = MJUM9BYTES;
 973 #endif
 974
 975         /* Prepare receive descriptors and buffers */
 976         for (i = 0; i < sc->rx_ring_inuse; ++i) {
 977                 int error;
 978
 979                 error = igb_init_rx_ring(&sc->rx_rings[i]);
 980                 if (error) {
 981                         if_printf(ifp, "Could not setup receive structures\n");
 982                         igb_stop(sc);
 983                         return;
 984                 }
 985         }
 986         igb_init_rx_unit(sc);
 987
 988         /* Enable VLAN support */
 989         if (ifp->if_capenable & IFCAP_VLAN_HWTAGGING)
 990                 igb_set_vlan(sc);
 991
 992         /* Don't lose promiscuous settings */
 993         igb_set_promisc(sc);
 994
 995         ifp->if_flags |= IFF_RUNNING;
 996         for (i = 0; i < sc->tx_ring_inuse; ++i) {
 997                 ifsq_clr_oactive(sc->tx_rings[i].ifsq);
 998                 ifsq_watchdog_start(&sc->tx_rings[i].tx_watchdog);
 999         }
1000
1001         igb_set_timer_cpuid(sc, polling);
1002         callout_reset_bycpu(&sc->timer, hz, igb_timer, sc, sc->timer_cpuid);
1003         e1000_clear_hw_cntrs_base_generic(&sc->hw);
1004
1005         /* This clears any pending interrupts */
1006         E1000_READ_REG(&sc->hw, E1000_ICR);
1007
1008         /*
1009          * Only enable interrupts if we are not polling, make sure
1010          * they are off otherwise.
1011          */
1012         if (polling) {
1013                 igb_disable_intr(sc);
1014         } else {
1015                 igb_enable_intr(sc);
1016                 E1000_WRITE_REG(&sc->hw, E1000_ICS, E1000_ICS_LSC);
1017         }
1018
1019         /* Set Energy Efficient Ethernet */
1020         if (sc->hw.phy.media_type == e1000_media_type_copper) {
1021                 if (sc->hw.mac.type == e1000_i354)
1022                         e1000_set_eee_i354(&sc->hw);
1023                 else
1024                         e1000_set_eee_i350(&sc->hw);
1025         }
1026 }
1027
1028 static void
1029 igb_media_status(struct ifnet *ifp, struct ifmediareq *ifmr)
1030 {
1031         struct igb_softc *sc = ifp->if_softc;
1032
1033         ASSERT_IFNET_SERIALIZED_ALL(ifp);
1034
1035         if ((ifp->if_flags & IFF_RUNNING) == 0)
1036                 sc->hw.mac.get_link_status = 1;
1037         igb_update_link_status(sc);
1038
1039         ifmr->ifm_status = IFM_AVALID;
1040         ifmr->ifm_active = IFM_ETHER;
1041
1042         if (!sc->link_active)
1043                 return;
1044
1045         ifmr->ifm_status |= IFM_ACTIVE;
1046
1047         switch (sc->link_speed) {
1048         case 10:
1049                 ifmr->ifm_active |= IFM_10_T;
1050                 break;
1051
1052         case 100:
1053                 /*
1054                  * Support for 100Mb SFP - these are Fiber
1055                  * but the media type appears as serdes
1056                  */
1057                 if (sc->hw.phy.media_type == e1000_media_type_internal_serdes)
1058                         ifmr->ifm_active |= IFM_100_FX;
1059                 else
1060                         ifmr->ifm_active |= IFM_100_TX;
1061                 break;
1062
1063         case 1000:
1064                 ifmr->ifm_active |= IFM_1000_T;
1065                 break;
1066         }
1067
1068         if (sc->link_duplex == FULL_DUPLEX)
1069                 ifmr->ifm_active |= IFM_FDX;
1070         else
1071                 ifmr->ifm_active |= IFM_HDX;
1072 }
1073
1074 static int
1075 igb_media_change(struct ifnet *ifp)
1076 {
1077         struct igb_softc *sc = ifp->if_softc;
1078         struct ifmedia *ifm = &sc->media;
1079
1080         ASSERT_IFNET_SERIALIZED_ALL(ifp);
1081
1082         if (IFM_TYPE(ifm->ifm_media) != IFM_ETHER)
1083                 return EINVAL;
1084
1085         switch (IFM_SUBTYPE(ifm->ifm_media)) {
1086         case IFM_AUTO:
1087                 sc->hw.mac.autoneg = DO_AUTO_NEG;
1088                 sc->hw.phy.autoneg_advertised = AUTONEG_ADV_DEFAULT;
1089                 break;
1090
1091         case IFM_1000_LX:
1092         case IFM_1000_SX:
1093         case IFM_1000_T:
1094                 sc->hw.mac.autoneg = DO_AUTO_NEG;
1095                 sc->hw.phy.autoneg_advertised = ADVERTISE_1000_FULL;
1096                 break;
1097
1098         case IFM_100_TX:
1099                 sc->hw.mac.autoneg = FALSE;
1100                 sc->hw.phy.autoneg_advertised = 0;
1101                 if ((ifm->ifm_media & IFM_GMASK) == IFM_FDX)
1102                         sc->hw.mac.forced_speed_duplex = ADVERTISE_100_FULL;
1103                 else
1104                         sc->hw.mac.forced_speed_duplex = ADVERTISE_100_HALF;
1105                 break;
1106
1107         case IFM_10_T:
1108                 sc->hw.mac.autoneg = FALSE;
1109                 sc->hw.phy.autoneg_advertised = 0;
1110                 if ((ifm->ifm_media & IFM_GMASK) == IFM_FDX)
1111                         sc->hw.mac.forced_speed_duplex = ADVERTISE_10_FULL;
1112                 else
1113                         sc->hw.mac.forced_speed_duplex = ADVERTISE_10_HALF;
1114                 break;
1115
1116         default:
1117                 if_printf(ifp, "Unsupported media type\n");
1118                 break;
1119         }
1120
1121         igb_init(sc);
1122
1123         return 0;
1124 }
1125
1126 static void
1127 igb_set_promisc(struct igb_softc *sc)
1128 {
1129         struct ifnet *ifp = &sc->arpcom.ac_if;
1130         struct e1000_hw *hw = &sc->hw;
1131         uint32_t reg;
1132
1133         if (sc->vf_ifp) {
1134                 e1000_promisc_set_vf(hw, e1000_promisc_enabled);
1135                 return;
1136         }
1137
1138         reg = E1000_READ_REG(hw, E1000_RCTL);
1139         if (ifp->if_flags & IFF_PROMISC) {
1140                 reg |= (E1000_RCTL_UPE | E1000_RCTL_MPE);
1141                 E1000_WRITE_REG(hw, E1000_RCTL, reg);
1142         } else if (ifp->if_flags & IFF_ALLMULTI) {
1143                 reg |= E1000_RCTL_MPE;
1144                 reg &= ~E1000_RCTL_UPE;
1145                 E1000_WRITE_REG(hw, E1000_RCTL, reg);
1146         }
1147 }
1148
1149 static void
1150 igb_disable_promisc(struct igb_softc *sc)
1151 {
1152         struct e1000_hw *hw = &sc->hw;
1153         struct ifnet *ifp = &sc->arpcom.ac_if;
1154         uint32_t reg;
1155         int mcnt = 0;
1156
1157         if (sc->vf_ifp) {
1158                 e1000_promisc_set_vf(hw, e1000_promisc_disabled);
1159                 return;
1160         }
1161         reg = E1000_READ_REG(hw, E1000_RCTL);
1162         reg &= ~E1000_RCTL_UPE;
1163         if (ifp->if_flags & IFF_ALLMULTI) {
1164                 mcnt = MAX_NUM_MULTICAST_ADDRESSES;
1165         } else {
1166                 struct  ifmultiaddr *ifma;
1167                 TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
1168                         if (ifma->ifma_addr->sa_family != AF_LINK)
1169                                 continue;
1170                         if (mcnt == MAX_NUM_MULTICAST_ADDRESSES)
1171                                 break;
1172                         mcnt++;
1173                 }
1174         }
1175         /* Don't disable if in MAX groups */
1176         if (mcnt < MAX_NUM_MULTICAST_ADDRESSES)
1177                 reg &= ~E1000_RCTL_MPE;
1178         E1000_WRITE_REG(hw, E1000_RCTL, reg);
1179 }
1180
1181 static void
1182 igb_set_multi(struct igb_softc *sc)
1183 {
1184         struct ifnet *ifp = &sc->arpcom.ac_if;
1185         struct ifmultiaddr *ifma;
1186         uint32_t reg_rctl = 0;
1187         uint8_t *mta;
1188         int mcnt = 0;
1189
1190         mta = sc->mta;
1191         bzero(mta, ETH_ADDR_LEN * MAX_NUM_MULTICAST_ADDRESSES);
1192
1193         TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
1194                 if (ifma->ifma_addr->sa_family != AF_LINK)
1195                         continue;
1196
1197                 if (mcnt == MAX_NUM_MULTICAST_ADDRESSES)
1198                         break;
1199
1200                 bcopy(LLADDR((struct sockaddr_dl *)ifma->ifma_addr),
1201                     &mta[mcnt * ETH_ADDR_LEN], ETH_ADDR_LEN);
1202                 mcnt++;
1203         }
1204
1205         if (mcnt >= MAX_NUM_MULTICAST_ADDRESSES) {
1206                 reg_rctl = E1000_READ_REG(&sc->hw, E1000_RCTL);
1207                 reg_rctl |= E1000_RCTL_MPE;
1208                 E1000_WRITE_REG(&sc->hw, E1000_RCTL, reg_rctl);
1209         } else {
1210                 e1000_update_mc_addr_list(&sc->hw, mta, mcnt);
1211         }
1212 }
1213
1214 static void
1215 igb_timer(void *xsc)
1216 {
1217         struct igb_softc *sc = xsc;
1218
1219         lwkt_serialize_enter(&sc->main_serialize);
1220
1221         igb_update_link_status(sc);
1222         igb_update_stats_counters(sc);
1223
1224         callout_reset_bycpu(&sc->timer, hz, igb_timer, sc, sc->timer_cpuid);
1225
1226         lwkt_serialize_exit(&sc->main_serialize);
1227 }
1228
1229 static void
1230 igb_update_link_status(struct igb_softc *sc)
1231 {
1232         struct ifnet *ifp = &sc->arpcom.ac_if;
1233         struct e1000_hw *hw = &sc->hw;
1234         uint32_t link_check, thstat, ctrl;
1235
1236         link_check = thstat = ctrl = 0;
1237
1238         /* Get the cached link value or read for real */
1239         switch (hw->phy.media_type) {
1240         case e1000_media_type_copper:
1241                 if (hw->mac.get_link_status) {
1242                         /* Do the work to read phy */
1243                         e1000_check_for_link(hw);
1244                         link_check = !hw->mac.get_link_status;
1245                 } else {
1246                         link_check = TRUE;
1247                 }
1248                 break;
1249
1250         case e1000_media_type_fiber:
1251                 e1000_check_for_link(hw);
1252                 link_check = E1000_READ_REG(hw, E1000_STATUS) & E1000_STATUS_LU;
1253                 break;
1254
1255         case e1000_media_type_internal_serdes:
1256                 e1000_check_for_link(hw);
1257                 link_check = hw->mac.serdes_has_link;
1258                 break;
1259
1260         /* VF device is type_unknown */
1261         case e1000_media_type_unknown:
1262                 e1000_check_for_link(hw);
1263                 link_check = !hw->mac.get_link_status;
1264                 /* Fall thru */
1265         default:
1266                 break;
1267         }
1268
1269         /* Check for thermal downshift or shutdown */
1270         if (hw->mac.type == e1000_i350) {
1271                 thstat = E1000_READ_REG(hw, E1000_THSTAT);
1272                 ctrl = E1000_READ_REG(hw, E1000_CTRL_EXT);
1273         }
1274
1275         /* Now we check if a transition has happened */
1276         if (link_check && sc->link_active == 0) {
1277                 e1000_get_speed_and_duplex(hw,
1278                     &sc->link_speed, &sc->link_duplex);
1279                 if (bootverbose) {
1280                         const char *flowctl;
1281
1282                         /* Get the flow control for display */
1283                         switch (hw->fc.current_mode) {
1284                         case e1000_fc_rx_pause:
1285                                 flowctl = "RX";
1286                                 break;
1287
1288                         case e1000_fc_tx_pause:
1289                                 flowctl = "TX";
1290                                 break;
1291
1292                         case e1000_fc_full:
1293                                 flowctl = "Full";
1294                                 break;
1295
1296                         default:
1297                                 flowctl = "None";
1298                                 break;
1299                         }
1300
1301                         if_printf(ifp, "Link is up %d Mbps %s, "
1302                             "Flow control: %s\n",
1303                             sc->link_speed,
1304                             sc->link_duplex == FULL_DUPLEX ?
1305                             "Full Duplex" : "Half Duplex",
1306                             flowctl);
1307                 }
1308                 sc->link_active = 1;
1309
1310                 ifp->if_baudrate = sc->link_speed * 1000000;
1311                 if ((ctrl & E1000_CTRL_EXT_LINK_MODE_GMII) &&
1312                     (thstat & E1000_THSTAT_LINK_THROTTLE))
1313                         if_printf(ifp, "Link: thermal downshift\n");
1314                 /* Delay Link Up for Phy update */
1315                 if ((hw->mac.type == e1000_i210 ||
1316                      hw->mac.type == e1000_i211) &&
1317                     hw->phy.id == I210_I_PHY_ID)
1318                         msec_delay(IGB_I210_LINK_DELAY);
1319                 /* This can sleep */
1320                 ifp->if_link_state = LINK_STATE_UP;
1321                 if_link_state_change(ifp);
1322         } else if (!link_check && sc->link_active == 1) {
1323                 ifp->if_baudrate = sc->link_speed = 0;
1324                 sc->link_duplex = 0;
1325                 if (bootverbose)
1326                         if_printf(ifp, "Link is Down\n");
1327                 if ((ctrl & E1000_CTRL_EXT_LINK_MODE_GMII) &&
1328                     (thstat & E1000_THSTAT_PWR_DOWN))
1329                         if_printf(ifp, "Link: thermal shutdown\n");
1330                 sc->link_active = 0;
1331                 /* This can sleep */
1332                 ifp->if_link_state = LINK_STATE_DOWN;
1333                 if_link_state_change(ifp);
1334         }
1335 }
1336
1337 static void
1338 igb_stop(struct igb_softc *sc)
1339 {
1340         struct ifnet *ifp = &sc->arpcom.ac_if;
1341         int i;
1342
1343         ASSERT_IFNET_SERIALIZED_ALL(ifp);
1344
1345         igb_disable_intr(sc);
1346
1347         callout_stop(&sc->timer);
1348
1349         ifp->if_flags &= ~IFF_RUNNING;
1350         for (i = 0; i < sc->tx_ring_cnt; ++i) {
1351                 ifsq_clr_oactive(sc->tx_rings[i].ifsq);
1352                 ifsq_watchdog_stop(&sc->tx_rings[i].tx_watchdog);
1353                 sc->tx_rings[i].tx_flags &= ~IGB_TXFLAG_ENABLED;
1354         }
1355
1356         e1000_reset_hw(&sc->hw);
1357         E1000_WRITE_REG(&sc->hw, E1000_WUC, 0);
1358
1359         e1000_led_off(&sc->hw);
1360         e1000_cleanup_led(&sc->hw);
1361
1362         for (i = 0; i < sc->tx_ring_cnt; ++i)
1363                 igb_free_tx_ring(&sc->tx_rings[i]);
1364         for (i = 0; i < sc->rx_ring_cnt; ++i)
1365                 igb_free_rx_ring(&sc->rx_rings[i]);
1366 }
1367
1368 static void
1369 igb_reset(struct igb_softc *sc)
1370 {
1371         struct ifnet *ifp = &sc->arpcom.ac_if;
1372         struct e1000_hw *hw = &sc->hw;
1373         struct e1000_fc_info *fc = &hw->fc;
1374         uint32_t pba = 0;
1375         uint16_t hwm;
1376
1377         /* Let the firmware know the OS is in control */
1378         igb_get_hw_control(sc);
1379
1380         /*
1381          * Packet Buffer Allocation (PBA)
1382          * Writing PBA sets the receive portion of the buffer
1383          * the remainder is used for the transmit buffer.
1384          */
1385         switch (hw->mac.type) {
1386         case e1000_82575:
1387                 pba = E1000_PBA_32K;
1388                 break;
1389
1390         case e1000_82576:
1391         case e1000_vfadapt:
1392                 pba = E1000_READ_REG(hw, E1000_RXPBS);
1393                 pba &= E1000_RXPBS_SIZE_MASK_82576;
1394                 break;
1395
1396         case e1000_82580:
1397         case e1000_i350:
1398         case e1000_i354:
1399         case e1000_vfadapt_i350:
1400                 pba = E1000_READ_REG(hw, E1000_RXPBS);
1401                 pba = e1000_rxpbs_adjust_82580(pba);
1402                 break;
1403
1404         case e1000_i210:
1405         case e1000_i211:
1406                 pba = E1000_PBA_34K;
1407                 break;
1408
1409         default:
1410                 break;
1411         }
1412
1413         /* Special needs in case of Jumbo frames */
1414         if (hw->mac.type == e1000_82575 && ifp->if_mtu > ETHERMTU) {
1415                 uint32_t tx_space, min_tx, min_rx;
1416
1417                 pba = E1000_READ_REG(hw, E1000_PBA);
1418                 tx_space = pba >> 16;
1419                 pba &= 0xffff;
1420
1421                 min_tx = (sc->max_frame_size +
1422                     sizeof(struct e1000_tx_desc) - ETHER_CRC_LEN) * 2;
1423                 min_tx = roundup2(min_tx, 1024);
1424                 min_tx >>= 10;
1425                 min_rx = sc->max_frame_size;
1426                 min_rx = roundup2(min_rx, 1024);
1427                 min_rx >>= 10;
1428                 if (tx_space < min_tx && (min_tx - tx_space) < pba) {
1429                         pba = pba - (min_tx - tx_space);
1430                         /*
1431                          * if short on rx space, rx wins
1432                          * and must trump tx adjustment
1433                          */
1434                         if (pba < min_rx)
1435                                 pba = min_rx;
1436                 }
1437                 E1000_WRITE_REG(hw, E1000_PBA, pba);
1438         }
1439
1440         /*
1441          * These parameters control the automatic generation (Tx) and
1442          * response (Rx) to Ethernet PAUSE frames.
1443          * - High water mark should allow for at least two frames to be
1444          *   received after sending an XOFF.
1445          * - Low water mark works best when it is very near the high water mark.
1446          *   This allows the receiver to restart by sending XON when it has
1447          *   drained a bit.
1448          */
1449         hwm = min(((pba << 10) * 9 / 10),
1450             ((pba << 10) - 2 * sc->max_frame_size));
1451
1452         if (hw->mac.type < e1000_82576) {
1453                 fc->high_water = hwm & 0xFFF8; /* 8-byte granularity */
1454                 fc->low_water = fc->high_water - 8;
1455         } else {
1456                 fc->high_water = hwm & 0xFFF0; /* 16-byte granularity */
1457                 fc->low_water = fc->high_water - 16;
1458         }
1459         fc->pause_time = IGB_FC_PAUSE_TIME;
1460         fc->send_xon = TRUE;
1461         fc->requested_mode = e1000_fc_default;
1462
1463         /* Issue a global reset */
1464         e1000_reset_hw(hw);
1465         E1000_WRITE_REG(hw, E1000_WUC, 0);
1466
1467         if (e1000_init_hw(hw) < 0)
1468                 if_printf(ifp, "Hardware Initialization Failed\n");
1469
1470         /* Setup DMA Coalescing */
1471         if (hw->mac.type > e1000_82580 && hw->mac.type != e1000_i211) {
1472                 uint32_t dmac;
1473                 uint32_t reg;
1474
1475                 if (sc->dma_coalesce == 0) {
1476                         /*
1477                          * Disabled
1478                          */
1479                         reg = E1000_READ_REG(hw, E1000_DMACR);
1480                         reg &= ~E1000_DMACR_DMAC_EN;
1481                         E1000_WRITE_REG(hw, E1000_DMACR, reg);
1482                         goto reset_out;
1483                 }
1484
1485                 /* Set starting thresholds */
1486                 E1000_WRITE_REG(hw, E1000_DMCTXTH, 0);
1487                 E1000_WRITE_REG(hw, E1000_DMCRTRH, 0);
1488
1489                 hwm = 64 * pba - sc->max_frame_size / 16;
1490                 if (hwm < 64 * (pba - 6))
1491                         hwm = 64 * (pba - 6);
1492                 reg = E1000_READ_REG(hw, E1000_FCRTC);
1493                 reg &= ~E1000_FCRTC_RTH_COAL_MASK;
1494                 reg |= ((hwm << E1000_FCRTC_RTH_COAL_SHIFT)
1495                     & E1000_FCRTC_RTH_COAL_MASK);
1496                 E1000_WRITE_REG(hw, E1000_FCRTC, reg);
1497
1498                 dmac = pba - sc->max_frame_size / 512;
1499                 if (dmac < pba - 10)
1500                         dmac = pba - 10;
1501                 reg = E1000_READ_REG(hw, E1000_DMACR);
1502                 reg &= ~E1000_DMACR_DMACTHR_MASK;
1503                 reg = ((dmac << E1000_DMACR_DMACTHR_SHIFT)
1504                     & E1000_DMACR_DMACTHR_MASK);
1505                 /* Transition to L0x or L1 if available.. */
1506                 reg |= (E1000_DMACR_DMAC_EN | E1000_DMACR_DMAC_LX_MASK);
1507                 /* timer = value in sc->dma_coalesce in 32usec intervals */
1508                 reg |= (sc->dma_coalesce >> 5);
1509                 E1000_WRITE_REG(hw, E1000_DMACR, reg);
1510
1511                 /* Set the interval before transition */
1512                 reg = E1000_READ_REG(hw, E1000_DMCTLX);
1513                 reg |= 0x80000004;
1514                 E1000_WRITE_REG(hw, E1000_DMCTLX, reg);
1515
1516                 /* Free space in tx packet buffer to wake from DMA coal */
1517                 E1000_WRITE_REG(hw, E1000_DMCTXTH,
1518                     (20480 - (2 * sc->max_frame_size)) >> 6);
1519
1520                 /* Make low power state decision controlled by DMA coal */
1521                 reg = E1000_READ_REG(hw, E1000_PCIEMISC);
1522                 reg &= ~E1000_PCIEMISC_LX_DECISION;
1523                 E1000_WRITE_REG(hw, E1000_PCIEMISC, reg);
1524                 if_printf(ifp, "DMA Coalescing enabled\n");
1525         } else if (hw->mac.type == e1000_82580) {
1526                 uint32_t reg = E1000_READ_REG(hw, E1000_PCIEMISC);
1527
1528                 E1000_WRITE_REG(hw, E1000_DMACR, 0);
1529                 E1000_WRITE_REG(hw, E1000_PCIEMISC,
1530                     reg & ~E1000_PCIEMISC_LX_DECISION);
1531         }
1532
1533 reset_out:
1534         E1000_WRITE_REG(&sc->hw, E1000_VET, ETHERTYPE_VLAN);
1535         e1000_get_phy_info(hw);
1536         e1000_check_for_link(hw);
1537 }
1538
1539 static void
1540 igb_setup_ifp(struct igb_softc *sc)
1541 {
1542         struct ifnet *ifp = &sc->arpcom.ac_if;
1543         int i;
1544
1545         ifp->if_softc = sc;
1546         ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
1547         ifp->if_init = igb_init;
1548         ifp->if_ioctl = igb_ioctl;
1549         ifp->if_start = igb_start;
1550         ifp->if_serialize = igb_serialize;
1551         ifp->if_deserialize = igb_deserialize;
1552         ifp->if_tryserialize = igb_tryserialize;
1553 #ifdef INVARIANTS
1554         ifp->if_serialize_assert = igb_serialize_assert;
1555 #endif
1556 #ifdef IFPOLL_ENABLE
1557         ifp->if_npoll = igb_npoll;
1558 #endif
1559
1560         ifp->if_nmbclusters = sc->rx_ring_cnt * sc->rx_rings[0].num_rx_desc;
1561
1562         ifq_set_maxlen(&ifp->if_snd, sc->tx_rings[0].num_tx_desc - 1);
1563         ifq_set_ready(&ifp->if_snd);
1564         ifq_set_subq_cnt(&ifp->if_snd, sc->tx_ring_cnt);
1565
1566         ifp->if_mapsubq = ifq_mapsubq_mask;
1567         ifq_set_subq_mask(&ifp->if_snd, 0);
1568
1569         ether_ifattach(ifp, sc->hw.mac.addr, NULL);
1570
1571         ifp->if_capabilities =
1572             IFCAP_HWCSUM | IFCAP_VLAN_HWTAGGING | IFCAP_VLAN_MTU | IFCAP_TSO;
1573         if (IGB_ENABLE_HWRSS(sc))
1574                 ifp->if_capabilities |= IFCAP_RSS;
1575         ifp->if_capenable = ifp->if_capabilities;
1576         ifp->if_hwassist = IGB_CSUM_FEATURES | CSUM_TSO;
1577
1578         /*
1579          * Tell the upper layer(s) we support long frames
1580          */
1581         ifp->if_data.ifi_hdrlen = sizeof(struct ether_vlan_header);
1582
1583         /* Setup TX rings and subqueues */
1584         for (i = 0; i < sc->tx_ring_cnt; ++i) {
1585                 struct ifaltq_subque *ifsq = ifq_get_subq(&ifp->if_snd, i);
1586                 struct igb_tx_ring *txr = &sc->tx_rings[i];
1587
1588                 ifsq_set_cpuid(ifsq, txr->tx_intr_cpuid);
1589                 ifsq_set_priv(ifsq, txr);
1590                 ifsq_set_hw_serialize(ifsq, &txr->tx_serialize);
1591                 txr->ifsq = ifsq;
1592
1593                 ifsq_watchdog_init(&txr->tx_watchdog, ifsq, igb_watchdog);
1594         }
1595
1596         /*
1597          * Specify the media types supported by this adapter and register
1598          * callbacks to update media and link information
1599          */
1600         ifmedia_init(&sc->media, IFM_IMASK, igb_media_change, igb_media_status);
1601         if (sc->hw.phy.media_type == e1000_media_type_fiber ||
1602             sc->hw.phy.media_type == e1000_media_type_internal_serdes) {
1603                 ifmedia_add(&sc->media, IFM_ETHER | IFM_1000_SX | IFM_FDX,
1604                     0, NULL);
1605                 ifmedia_add(&sc->media, IFM_ETHER | IFM_1000_SX, 0, NULL);
1606         } else {
1607                 ifmedia_add(&sc->media, IFM_ETHER | IFM_10_T, 0, NULL);
1608                 ifmedia_add(&sc->media, IFM_ETHER | IFM_10_T | IFM_FDX,
1609                     0, NULL);
1610                 ifmedia_add(&sc->media, IFM_ETHER | IFM_100_TX, 0, NULL);
1611                 ifmedia_add(&sc->media, IFM_ETHER | IFM_100_TX | IFM_FDX,
1612                     0, NULL);
1613                 if (sc->hw.phy.type != e1000_phy_ife) {
1614                         ifmedia_add(&sc->media,
1615                             IFM_ETHER | IFM_1000_T | IFM_FDX, 0, NULL);
1616                         ifmedia_add(&sc->media,
1617                             IFM_ETHER | IFM_1000_T, 0, NULL);
1618                 }
1619         }
1620         ifmedia_add(&sc->media, IFM_ETHER | IFM_AUTO, 0, NULL);
1621         ifmedia_set(&sc->media, IFM_ETHER | IFM_AUTO);
1622 }
1623
1624 static void
1625 igb_add_sysctl(struct igb_softc *sc)
1626 {
1627         struct sysctl_ctx_list *ctx;
1628         struct sysctl_oid *tree;
1629         char node[32];
1630         int i;
1631
1632         ctx = device_get_sysctl_ctx(sc->dev);
1633         tree = device_get_sysctl_tree(sc->dev);
1634         SYSCTL_ADD_INT(ctx, SYSCTL_CHILDREN(tree),
1635             OID_AUTO, "rxr", CTLFLAG_RD, &sc->rx_ring_cnt, 0, "# of RX rings");
1636         SYSCTL_ADD_INT(ctx, SYSCTL_CHILDREN(tree),
1637             OID_AUTO, "rxr_inuse", CTLFLAG_RD, &sc->rx_ring_inuse, 0,
1638             "# of RX rings used");
1639         SYSCTL_ADD_INT(ctx, SYSCTL_CHILDREN(tree),
1640             OID_AUTO, "txr", CTLFLAG_RD, &sc->tx_ring_cnt, 0, "# of TX rings");
1641         SYSCTL_ADD_INT(ctx, SYSCTL_CHILDREN(tree),
1642             OID_AUTO, "txr_inuse", CTLFLAG_RD, &sc->tx_ring_inuse, 0,
1643             "# of TX rings used");
1644         SYSCTL_ADD_INT(ctx, SYSCTL_CHILDREN(tree),
1645             OID_AUTO, "rxd", CTLFLAG_RD, &sc->rx_rings[0].num_rx_desc, 0,
1646             "# of RX descs");
1647         SYSCTL_ADD_INT(ctx, SYSCTL_CHILDREN(tree),
1648             OID_AUTO, "txd", CTLFLAG_RD, &sc->tx_rings[0].num_tx_desc, 0,
1649             "# of TX descs");
1650
1651         if (sc->intr_type != PCI_INTR_TYPE_MSIX) {
1652                 SYSCTL_ADD_PROC(ctx, SYSCTL_CHILDREN(tree),
1653                     OID_AUTO, "intr_rate", CTLTYPE_INT | CTLFLAG_RW,
1654                     sc, 0, igb_sysctl_intr_rate, "I", "interrupt rate");
1655         } else {
1656                 for (i = 0; i < sc->msix_cnt; ++i) {
1657                         struct igb_msix_data *msix = &sc->msix_data[i];
1658
1659                         ksnprintf(node, sizeof(node), "msix%d_rate", i);
1660                         SYSCTL_ADD_PROC(ctx, SYSCTL_CHILDREN(tree),
1661                             OID_AUTO, node, CTLTYPE_INT | CTLFLAG_RW,
1662                             msix, 0, igb_sysctl_msix_rate, "I",
1663                             msix->msix_rate_desc);
1664                 }
1665         }
1666
1667         SYSCTL_ADD_PROC(ctx, SYSCTL_CHILDREN(tree),
1668             OID_AUTO, "tx_intr_nsegs", CTLTYPE_INT | CTLFLAG_RW,
1669             sc, 0, igb_sysctl_tx_intr_nsegs, "I",
1670             "# of segments per TX interrupt");
1671
1672         SYSCTL_ADD_PROC(ctx, SYSCTL_CHILDREN(tree),
1673             OID_AUTO, "tx_wreg_nsegs", CTLTYPE_INT | CTLFLAG_RW,
1674             sc, 0, igb_sysctl_tx_wreg_nsegs, "I",
1675             "# of segments sent before write to hardware register");
1676
1677         SYSCTL_ADD_PROC(ctx, SYSCTL_CHILDREN(tree),
1678             OID_AUTO, "rx_wreg_nsegs", CTLTYPE_INT | CTLFLAG_RW,
1679             sc, 0, igb_sysctl_rx_wreg_nsegs, "I",
1680             "# of segments received before write to hardware register");
1681
1682 #ifdef IFPOLL_ENABLE
1683         SYSCTL_ADD_PROC(ctx, SYSCTL_CHILDREN(tree),
1684             OID_AUTO, "npoll_rxoff", CTLTYPE_INT|CTLFLAG_RW,
1685             sc, 0, igb_sysctl_npoll_rxoff, "I", "NPOLLING RX cpu offset");
1686         SYSCTL_ADD_PROC(ctx, SYSCTL_CHILDREN(tree),
1687             OID_AUTO, "npoll_txoff", CTLTYPE_INT|CTLFLAG_RW,
1688             sc, 0, igb_sysctl_npoll_txoff, "I", "NPOLLING TX cpu offset");
1689 #endif
1690
1691 #ifdef IGB_RSS_DEBUG
1692         SYSCTL_ADD_INT(ctx, SYSCTL_CHILDREN(tree),
1693             OID_AUTO, "rss_debug", CTLFLAG_RW, &sc->rss_debug, 0,
1694             "RSS debug level");
1695         for (i = 0; i < sc->rx_ring_cnt; ++i) {
1696                 ksnprintf(node, sizeof(node), "rx%d_pkt", i);
1697                 SYSCTL_ADD_ULONG(ctx,
1698                     SYSCTL_CHILDREN(tree), OID_AUTO, node,
1699                     CTLFLAG_RW, &sc->rx_rings[i].rx_packets, "RXed packets");
1700         }
1701 #endif
1702 #ifdef IGB_TSS_DEBUG
1703         for  (i = 0; i < sc->tx_ring_cnt; ++i) {
1704                 ksnprintf(node, sizeof(node), "tx%d_pkt", i);
1705                 SYSCTL_ADD_ULONG(ctx,
1706                     SYSCTL_CHILDREN(tree), OID_AUTO, node,
1707                     CTLFLAG_RW, &sc->tx_rings[i].tx_packets, "TXed packets");
1708         }
1709 #endif
1710 }
1711
1712 static int
1713 igb_alloc_rings(struct igb_softc *sc)
1714 {
1715         int error, i;
1716
1717         /*
1718          * Create top level busdma tag
1719          */
1720         error = bus_dma_tag_create(NULL, 1, 0,
1721             BUS_SPACE_MAXADDR, BUS_SPACE_MAXADDR, NULL, NULL,
1722             BUS_SPACE_MAXSIZE_32BIT, 0, BUS_SPACE_MAXSIZE_32BIT, 0,
1723             &sc->parent_tag);
1724         if (error) {
1725                 device_printf(sc->dev, "could not create top level DMA tag\n");
1726                 return error;
1727         }
1728
1729         /*
1730          * Allocate TX descriptor rings and buffers
1731          */
1732         sc->tx_rings = kmalloc_cachealign(
1733             sizeof(struct igb_tx_ring) * sc->tx_ring_cnt,
1734             M_DEVBUF, M_WAITOK | M_ZERO);
1735         for (i = 0; i < sc->tx_ring_cnt; ++i) {
1736                 struct igb_tx_ring *txr = &sc->tx_rings[i];
1737
1738                 /* Set up some basics */
1739                 txr->sc = sc;
1740                 txr->me = i;
1741                 lwkt_serialize_init(&txr->tx_serialize);
1742
1743                 error = igb_create_tx_ring(txr);
1744                 if (error)
1745                         return error;
1746         }
1747
1748         /*
1749          * Allocate RX descriptor rings and buffers
1750          */
1751         sc->rx_rings = kmalloc_cachealign(
1752             sizeof(struct igb_rx_ring) * sc->rx_ring_cnt,
1753             M_DEVBUF, M_WAITOK | M_ZERO);
1754         for (i = 0; i < sc->rx_ring_cnt; ++i) {
1755                 struct igb_rx_ring *rxr = &sc->rx_rings[i];
1756
1757                 /* Set up some basics */
1758                 rxr->sc = sc;
1759                 rxr->me = i;
1760                 lwkt_serialize_init(&rxr->rx_serialize);
1761
1762                 error = igb_create_rx_ring(rxr);
1763                 if (error)
1764                         return error;
1765         }
1766
1767         return 0;
1768 }
1769
1770 static void
1771 igb_free_rings(struct igb_softc *sc)
1772 {
1773         int i;
1774
1775         if (sc->tx_rings != NULL) {
1776                 for (i = 0; i < sc->tx_ring_cnt; ++i) {
1777                         struct igb_tx_ring *txr = &sc->tx_rings[i];
1778
1779                         igb_destroy_tx_ring(txr, txr->num_tx_desc);
1780                 }
1781                 kfree(sc->tx_rings, M_DEVBUF);
1782         }
1783
1784         if (sc->rx_rings != NULL) {
1785                 for (i = 0; i < sc->rx_ring_cnt; ++i) {
1786                         struct igb_rx_ring *rxr = &sc->rx_rings[i];
1787
1788                         igb_destroy_rx_ring(rxr, rxr->num_rx_desc);
1789                 }
1790                 kfree(sc->rx_rings, M_DEVBUF);
1791         }
1792 }
1793
1794 static int
1795 igb_create_tx_ring(struct igb_tx_ring *txr)
1796 {
1797         int tsize, error, i, ntxd;
1798
1799         /*
1800          * Validate number of transmit descriptors. It must not exceed
1801          * hardware maximum, and must be multiple of IGB_DBA_ALIGN.
1802          */
1803         ntxd = device_getenv_int(txr->sc->dev, "txd", igb_txd);
1804         if ((ntxd * sizeof(struct e1000_tx_desc)) % IGB_DBA_ALIGN != 0 ||
1805             ntxd > IGB_MAX_TXD || ntxd < IGB_MIN_TXD) {
1806                 device_printf(txr->sc->dev,
1807                     "Using %d TX descriptors instead of %d!\n",
1808                     IGB_DEFAULT_TXD, ntxd);
1809                 txr->num_tx_desc = IGB_DEFAULT_TXD;
1810         } else {
1811                 txr->num_tx_desc = ntxd;
1812         }
1813
1814         /*
1815          * Allocate TX descriptor ring
1816          */
1817         tsize = roundup2(txr->num_tx_desc * sizeof(union e1000_adv_tx_desc),
1818             IGB_DBA_ALIGN);
1819         txr->txdma.dma_vaddr = bus_dmamem_coherent_any(txr->sc->parent_tag,
1820             IGB_DBA_ALIGN, tsize, BUS_DMA_WAITOK,
1821             &txr->txdma.dma_tag, &txr->txdma.dma_map, &txr->txdma.dma_paddr);
1822         if (txr->txdma.dma_vaddr == NULL) {
1823                 device_printf(txr->sc->dev,
1824                     "Unable to allocate TX Descriptor memory\n");
1825                 return ENOMEM;
1826         }
1827         txr->tx_base = txr->txdma.dma_vaddr;
1828         bzero(txr->tx_base, tsize);
1829
1830         tsize = __VM_CACHELINE_ALIGN(
1831             sizeof(struct igb_tx_buf) * txr->num_tx_desc);
1832         txr->tx_buf = kmalloc_cachealign(tsize, M_DEVBUF, M_WAITOK | M_ZERO);
1833
1834         /*
1835          * Allocate TX head write-back buffer
1836          */
1837         txr->tx_hdr = bus_dmamem_coherent_any(txr->sc->parent_tag,
1838             __VM_CACHELINE_SIZE, __VM_CACHELINE_SIZE, BUS_DMA_WAITOK,
1839             &txr->tx_hdr_dtag, &txr->tx_hdr_dmap, &txr->tx_hdr_paddr);
1840         if (txr->tx_hdr == NULL) {
1841                 device_printf(txr->sc->dev,
1842                     "Unable to allocate TX head write-back buffer\n");
1843                 return ENOMEM;
1844         }
1845
1846         /*
1847          * Create DMA tag for TX buffers
1848          */
1849         error = bus_dma_tag_create(txr->sc->parent_tag,
1850             1, 0,               /* alignment, bounds */
1851             BUS_SPACE_MAXADDR,  /* lowaddr */
1852             BUS_SPACE_MAXADDR,  /* highaddr */
1853             NULL, NULL,         /* filter, filterarg */
1854             IGB_TSO_SIZE,       /* maxsize */
1855             IGB_MAX_SCATTER,    /* nsegments */
1856             PAGE_SIZE,          /* maxsegsize */
1857             BUS_DMA_WAITOK | BUS_DMA_ALLOCNOW |
1858             BUS_DMA_ONEBPAGE,   /* flags */
1859             &txr->tx_tag);
1860         if (error) {
1861                 device_printf(txr->sc->dev, "Unable to allocate TX DMA tag\n");
1862                 kfree(txr->tx_buf, M_DEVBUF);
1863                 txr->tx_buf = NULL;
1864                 return error;
1865         }
1866
1867         /*
1868          * Create DMA maps for TX buffers
1869          */
1870         for (i = 0; i < txr->num_tx_desc; ++i) {
1871                 struct igb_tx_buf *txbuf = &txr->tx_buf[i];
1872
1873                 error = bus_dmamap_create(txr->tx_tag,
1874                     BUS_DMA_WAITOK | BUS_DMA_ONEBPAGE, &txbuf->map);
1875                 if (error) {
1876                         device_printf(txr->sc->dev,
1877                             "Unable to create TX DMA map\n");
1878                         igb_destroy_tx_ring(txr, i);
1879                         return error;
1880                 }
1881         }
1882
1883         if (txr->sc->hw.mac.type == e1000_82575)
1884                 txr->tx_flags |= IGB_TXFLAG_TSO_IPLEN0;
1885
1886         /*
1887          * Initialize various watermark
1888          */
1889         txr->spare_desc = IGB_TX_SPARE;
1890         txr->intr_nsegs = txr->num_tx_desc / 16;
1891         txr->wreg_nsegs = IGB_DEF_TXWREG_NSEGS;
1892         txr->oact_hi_desc = txr->num_tx_desc / 2;
1893         txr->oact_lo_desc = txr->num_tx_desc / 8;
1894         if (txr->oact_lo_desc > IGB_TX_OACTIVE_MAX)
1895                 txr->oact_lo_desc = IGB_TX_OACTIVE_MAX;
1896         if (txr->oact_lo_desc < txr->spare_desc + IGB_TX_RESERVED)
1897                 txr->oact_lo_desc = txr->spare_desc + IGB_TX_RESERVED;
1898
1899         return 0;
1900 }
1901
1902 static void
1903 igb_free_tx_ring(struct igb_tx_ring *txr)
1904 {
1905         int i;
1906
1907         for (i = 0; i < txr->num_tx_desc; ++i) {
1908                 struct igb_tx_buf *txbuf = &txr->tx_buf[i];
1909
1910                 if (txbuf->m_head != NULL) {
1911                         bus_dmamap_unload(txr->tx_tag, txbuf->map);
1912                         m_freem(txbuf->m_head);
1913                         txbuf->m_head = NULL;
1914                 }
1915         }
1916 }
1917
1918 static void
1919 igb_destroy_tx_ring(struct igb_tx_ring *txr, int ndesc)
1920 {
1921         int i;
1922
1923         if (txr->txdma.dma_vaddr != NULL) {
1924                 bus_dmamap_unload(txr->txdma.dma_tag, txr->txdma.dma_map);
1925                 bus_dmamem_free(txr->txdma.dma_tag, txr->txdma.dma_vaddr,
1926                     txr->txdma.dma_map);
1927                 bus_dma_tag_destroy(txr->txdma.dma_tag);
1928                 txr->txdma.dma_vaddr = NULL;
1929         }
1930
1931         if (txr->tx_hdr != NULL) {
1932                 bus_dmamap_unload(txr->tx_hdr_dtag, txr->tx_hdr_dmap);
1933                 bus_dmamem_free(txr->tx_hdr_dtag, txr->tx_hdr,
1934                     txr->tx_hdr_dmap);
1935                 bus_dma_tag_destroy(txr->tx_hdr_dtag);
1936                 txr->tx_hdr = NULL;
1937         }
1938
1939         if (txr->tx_buf == NULL)
1940                 return;
1941
1942         for (i = 0; i < ndesc; ++i) {
1943                 struct igb_tx_buf *txbuf = &txr->tx_buf[i];
1944
1945                 KKASSERT(txbuf->m_head == NULL);
1946                 bus_dmamap_destroy(txr->tx_tag, txbuf->map);
1947         }
1948         bus_dma_tag_destroy(txr->tx_tag);
1949
1950         kfree(txr->tx_buf, M_DEVBUF);
1951         txr->tx_buf = NULL;
1952 }
1953
1954 static void
1955 igb_init_tx_ring(struct igb_tx_ring *txr)
1956 {
1957         /* Clear the old descriptor contents */
1958         bzero(txr->tx_base,
1959             sizeof(union e1000_adv_tx_desc) * txr->num_tx_desc);
1960
1961         /* Clear TX head write-back buffer */
1962         *(txr->tx_hdr) = 0;
1963
1964         /* Reset indices */
1965         txr->next_avail_desc = 0;
1966         txr->next_to_clean = 0;
1967         txr->tx_nsegs = 0;
1968
1969         /* Set number of descriptors available */
1970         txr->tx_avail = txr->num_tx_desc;
1971
1972         /* Enable this TX ring */
1973         txr->tx_flags |= IGB_TXFLAG_ENABLED;
1974 }
1975
1976 static void
1977 igb_init_tx_unit(struct igb_softc *sc)
1978 {
1979         struct e1000_hw *hw = &sc->hw;
1980         uint32_t tctl;
1981         int i;
1982
1983         /* Setup the Tx Descriptor Rings */
1984         for (i = 0; i < sc->tx_ring_inuse; ++i) {
1985                 struct igb_tx_ring *txr = &sc->tx_rings[i];
1986                 uint64_t bus_addr = txr->txdma.dma_paddr;
1987                 uint64_t hdr_paddr = txr->tx_hdr_paddr;
1988                 uint32_t txdctl = 0;
1989                 uint32_t dca_txctrl;
1990
1991                 E1000_WRITE_REG(hw, E1000_TDLEN(i),
1992                     txr->num_tx_desc * sizeof(struct e1000_tx_desc));
1993                 E1000_WRITE_REG(hw, E1000_TDBAH(i),
1994                     (uint32_t)(bus_addr >> 32));
1995                 E1000_WRITE_REG(hw, E1000_TDBAL(i),
1996                     (uint32_t)bus_addr);
1997
1998                 /* Setup the HW Tx Head and Tail descriptor pointers */
1999                 E1000_WRITE_REG(hw, E1000_TDT(i), 0);
2000                 E1000_WRITE_REG(hw, E1000_TDH(i), 0);
2001
2002                 dca_txctrl = E1000_READ_REG(hw, E1000_DCA_TXCTRL(i));
2003                 dca_txctrl &= ~E1000_DCA_TXCTRL_TX_WB_RO_EN;
2004                 E1000_WRITE_REG(hw, E1000_DCA_TXCTRL(i), dca_txctrl);
2005
2006                 /*
2007                  * Don't set WB_on_EITR:
2008                  * - 82575 does not have it
2009                  * - It almost has no effect on 82576, see:
2010                  *   82576 specification update errata #26
2011                  * - It causes unnecessary bus traffic
2012                  */
2013                 E1000_WRITE_REG(hw, E1000_TDWBAH(i),
2014                     (uint32_t)(hdr_paddr >> 32));
2015                 E1000_WRITE_REG(hw, E1000_TDWBAL(i),
2016                     ((uint32_t)hdr_paddr) | E1000_TX_HEAD_WB_ENABLE);
2017
2018                 /*
2019                  * WTHRESH is ignored by the hardware, since header
2020                  * write back mode is used.
2021                  */
2022                 txdctl |= IGB_TX_PTHRESH;
2023                 txdctl |= IGB_TX_HTHRESH << 8;
2024                 txdctl |= IGB_TX_WTHRESH << 16;
2025                 txdctl |= E1000_TXDCTL_QUEUE_ENABLE;
2026                 E1000_WRITE_REG(hw, E1000_TXDCTL(i), txdctl);
2027         }
2028
2029         if (sc->vf_ifp)
2030                 return;
2031
2032         e1000_config_collision_dist(hw);
2033
2034         /* Program the Transmit Control Register */
2035         tctl = E1000_READ_REG(hw, E1000_TCTL);
2036         tctl &= ~E1000_TCTL_CT;
2037         tctl |= (E1000_TCTL_PSP | E1000_TCTL_RTLC | E1000_TCTL_EN |
2038             (E1000_COLLISION_THRESHOLD << E1000_CT_SHIFT));
2039
2040         /* This write will effectively turn on the transmit unit. */
2041         E1000_WRITE_REG(hw, E1000_TCTL, tctl);
2042 }
2043
2044 static boolean_t
2045 igb_txcsum_ctx(struct igb_tx_ring *txr, struct mbuf *mp)
2046 {
2047         struct e1000_adv_tx_context_desc *TXD;
2048         uint32_t vlan_macip_lens, type_tucmd_mlhl, mss_l4len_idx;
2049         int ehdrlen, ctxd, ip_hlen = 0;
2050         boolean_t offload = TRUE;
2051
2052         if ((mp->m_pkthdr.csum_flags & IGB_CSUM_FEATURES) == 0)
2053                 offload = FALSE;
2054
2055         vlan_macip_lens = type_tucmd_mlhl = mss_l4len_idx = 0;
2056
2057         ctxd = txr->next_avail_desc;
2058         TXD = (struct e1000_adv_tx_context_desc *)&txr->tx_base[ctxd];
2059
2060         /*
2061          * In advanced descriptors the vlan tag must
2062          * be placed into the context descriptor, thus
2063          * we need to be here just for that setup.
2064          */
2065         if (mp->m_flags & M_VLANTAG) {
2066                 uint16_t vlantag;
2067
2068                 vlantag = htole16(mp->m_pkthdr.ether_vlantag);
2069                 vlan_macip_lens |= (vlantag << E1000_ADVTXD_VLAN_SHIFT);
2070         } else if (!offload) {
2071                 return FALSE;
2072         }
2073
2074         ehdrlen = mp->m_pkthdr.csum_lhlen;
2075         KASSERT(ehdrlen > 0, ("invalid ether hlen"));
2076
2077         /* Set the ether header length */
2078         vlan_macip_lens |= ehdrlen << E1000_ADVTXD_MACLEN_SHIFT;
2079         if (mp->m_pkthdr.csum_flags & CSUM_IP) {
2080                 type_tucmd_mlhl |= E1000_ADVTXD_TUCMD_IPV4;
2081                 ip_hlen = mp->m_pkthdr.csum_iphlen;
2082                 KASSERT(ip_hlen > 0, ("invalid ip hlen"));
2083         }
2084         vlan_macip_lens |= ip_hlen;
2085
2086         type_tucmd_mlhl |= E1000_ADVTXD_DCMD_DEXT | E1000_ADVTXD_DTYP_CTXT;
2087         if (mp->m_pkthdr.csum_flags & CSUM_TCP)
2088                 type_tucmd_mlhl |= E1000_ADVTXD_TUCMD_L4T_TCP;
2089         else if (mp->m_pkthdr.csum_flags & CSUM_UDP)
2090                 type_tucmd_mlhl |= E1000_ADVTXD_TUCMD_L4T_UDP;
2091
2092         /*
2093          * 82575 needs the TX context index added; the queue
2094          * index is used as TX context index here.
2095          */
2096         if (txr->sc->hw.mac.type == e1000_82575)
2097                 mss_l4len_idx = txr->me << 4;
2098
2099         /* Now copy bits into descriptor */
2100         TXD->vlan_macip_lens = htole32(vlan_macip_lens);
2101         TXD->type_tucmd_mlhl = htole32(type_tucmd_mlhl);
2102         TXD->seqnum_seed = htole32(0);
2103         TXD->mss_l4len_idx = htole32(mss_l4len_idx);
2104
2105         /* We've consumed the first desc, adjust counters */
2106         if (++ctxd == txr->num_tx_desc)
2107                 ctxd = 0;
2108         txr->next_avail_desc = ctxd;
2109         --txr->tx_avail;
2110
2111         return offload;
2112 }
2113
2114 static void
2115 igb_txeof(struct igb_tx_ring *txr)
2116 {
2117         int first, hdr, avail;
2118
2119         if (txr->tx_avail == txr->num_tx_desc)
2120                 return;
2121
2122         first = txr->next_to_clean;
2123         hdr = *(txr->tx_hdr);
2124
2125         if (first == hdr)
2126                 return;
2127
2128         avail = txr->tx_avail;
2129         while (first != hdr) {
2130                 struct igb_tx_buf *txbuf = &txr->tx_buf[first];
2131
2132                 ++avail;
2133                 if (txbuf->m_head) {
2134                         bus_dmamap_unload(txr->tx_tag, txbuf->map);
2135                         m_freem(txbuf->m_head);
2136                         txbuf->m_head = NULL;
2137                 }
2138                 if (++first == txr->num_tx_desc)
2139                         first = 0;
2140         }
2141         txr->next_to_clean = first;
2142         txr->tx_avail = avail;
2143
2144         /*
2145          * If we have a minimum free, clear OACTIVE
2146          * to tell the stack that it is OK to send packets.
2147          */
2148         if (IGB_IS_NOT_OACTIVE(txr)) {
2149                 ifsq_clr_oactive(txr->ifsq);
2150
2151                 /*
2152                  * We have enough TX descriptors, turn off
2153                  * the watchdog.  We allow small amount of
2154                  * packets (roughly intr_nsegs) pending on
2155                  * the transmit ring.
2156                  */
2157                 txr->tx_watchdog.wd_timer = 0;
2158         }
2159 }
2160
2161 static int
2162 igb_create_rx_ring(struct igb_rx_ring *rxr)
2163 {
2164         int rsize, i, error, nrxd;
2165
2166         /*
2167          * Validate number of receive descriptors. It must not exceed
2168          * hardware maximum, and must be multiple of IGB_DBA_ALIGN.
2169          */
2170         nrxd = device_getenv_int(rxr->sc->dev, "rxd", igb_rxd);
2171         if ((nrxd * sizeof(struct e1000_rx_desc)) % IGB_DBA_ALIGN != 0 ||
2172             nrxd > IGB_MAX_RXD || nrxd < IGB_MIN_RXD) {
2173                 device_printf(rxr->sc->dev,
2174                     "Using %d RX descriptors instead of %d!\n",
2175                     IGB_DEFAULT_RXD, nrxd);
2176                 rxr->num_rx_desc = IGB_DEFAULT_RXD;
2177         } else {
2178                 rxr->num_rx_desc = nrxd;
2179         }
2180
2181         /*
2182          * Allocate RX descriptor ring
2183          */
2184         rsize = roundup2(rxr->num_rx_desc * sizeof(union e1000_adv_rx_desc),
2185             IGB_DBA_ALIGN);
2186         rxr->rxdma.dma_vaddr = bus_dmamem_coherent_any(rxr->sc->parent_tag,
2187             IGB_DBA_ALIGN, rsize, BUS_DMA_WAITOK,
2188             &rxr->rxdma.dma_tag, &rxr->rxdma.dma_map,
2189             &rxr->rxdma.dma_paddr);
2190         if (rxr->rxdma.dma_vaddr == NULL) {
2191                 device_printf(rxr->sc->dev,
2192                     "Unable to allocate RxDescriptor memory\n");
2193                 return ENOMEM;
2194         }
2195         rxr->rx_base = rxr->rxdma.dma_vaddr;
2196         bzero(rxr->rx_base, rsize);
2197
2198         rsize = __VM_CACHELINE_ALIGN(
2199             sizeof(struct igb_rx_buf) * rxr->num_rx_desc);
2200         rxr->rx_buf = kmalloc_cachealign(rsize, M_DEVBUF, M_WAITOK | M_ZERO);
2201
2202         /*
2203          * Create DMA tag for RX buffers
2204          */
2205         error = bus_dma_tag_create(rxr->sc->parent_tag,
2206             1, 0,               /* alignment, bounds */
2207             BUS_SPACE_MAXADDR,  /* lowaddr */
2208             BUS_SPACE_MAXADDR,  /* highaddr */
2209             NULL, NULL,         /* filter, filterarg */
2210             MCLBYTES,           /* maxsize */
2211             1,                  /* nsegments */
2212             MCLBYTES,           /* maxsegsize */
2213             BUS_DMA_WAITOK | BUS_DMA_ALLOCNOW, /* flags */
2214             &rxr->rx_tag);
2215         if (error) {
2216                 device_printf(rxr->sc->dev,
2217                     "Unable to create RX payload DMA tag\n");
2218                 kfree(rxr->rx_buf, M_DEVBUF);
2219                 rxr->rx_buf = NULL;
2220                 return error;
2221         }
2222
2223         /*
2224          * Create spare DMA map for RX buffers
2225          */
2226         error = bus_dmamap_create(rxr->rx_tag, BUS_DMA_WAITOK,
2227             &rxr->rx_sparemap);
2228         if (error) {
2229                 device_printf(rxr->sc->dev,
2230                     "Unable to create spare RX DMA maps\n");
2231                 bus_dma_tag_destroy(rxr->rx_tag);
2232                 kfree(rxr->rx_buf, M_DEVBUF);
2233                 rxr->rx_buf = NULL;
2234                 return error;
2235         }
2236
2237         /*
2238          * Create DMA maps for RX buffers
2239          */
2240         for (i = 0; i < rxr->num_rx_desc; i++) {
2241                 struct igb_rx_buf *rxbuf = &rxr->rx_buf[i];
2242
2243                 error = bus_dmamap_create(rxr->rx_tag,
2244                     BUS_DMA_WAITOK, &rxbuf->map);
2245                 if (error) {
2246                         device_printf(rxr->sc->dev,
2247                             "Unable to create RX DMA maps\n");
2248                         igb_destroy_rx_ring(rxr, i);
2249                         return error;
2250                 }
2251         }
2252
2253         /*
2254          * Initialize various watermark
2255          */
2256         rxr->wreg_nsegs = IGB_DEF_RXWREG_NSEGS;
2257
2258         return 0;
2259 }
2260
2261 static void
2262 igb_free_rx_ring(struct igb_rx_ring *rxr)
2263 {
2264         int i;
2265
2266         for (i = 0; i < rxr->num_rx_desc; ++i) {
2267                 struct igb_rx_buf *rxbuf = &rxr->rx_buf[i];
2268
2269                 if (rxbuf->m_head != NULL) {
2270                         bus_dmamap_unload(rxr->rx_tag, rxbuf->map);
2271                         m_freem(rxbuf->m_head);
2272                         rxbuf->m_head = NULL;
2273                 }
2274         }
2275
2276         if (rxr->fmp != NULL)
2277                 m_freem(rxr->fmp);
2278         rxr->fmp = NULL;
2279         rxr->lmp = NULL;
2280 }
2281
2282 static void
2283 igb_destroy_rx_ring(struct igb_rx_ring *rxr, int ndesc)
2284 {
2285         int i;
2286
2287         if (rxr->rxdma.dma_vaddr != NULL) {
2288                 bus_dmamap_unload(rxr->rxdma.dma_tag, rxr->rxdma.dma_map);
2289                 bus_dmamem_free(rxr->rxdma.dma_tag, rxr->rxdma.dma_vaddr,
2290                     rxr->rxdma.dma_map);
2291                 bus_dma_tag_destroy(rxr->rxdma.dma_tag);
2292                 rxr->rxdma.dma_vaddr = NULL;
2293         }
2294
2295         if (rxr->rx_buf == NULL)
2296                 return;
2297
2298         for (i = 0; i < ndesc; ++i) {
2299                 struct igb_rx_buf *rxbuf = &rxr->rx_buf[i];
2300
2301                 KKASSERT(rxbuf->m_head == NULL);
2302                 bus_dmamap_destroy(rxr->rx_tag, rxbuf->map);
2303         }
2304         bus_dmamap_destroy(rxr->rx_tag, rxr->rx_sparemap);
2305         bus_dma_tag_destroy(rxr->rx_tag);
2306
2307         kfree(rxr->rx_buf, M_DEVBUF);
2308         rxr->rx_buf = NULL;
2309 }
2310
2311 static void
2312 igb_setup_rxdesc(union e1000_adv_rx_desc *rxd, const struct igb_rx_buf *rxbuf)
2313 {
2314         rxd->read.pkt_addr = htole64(rxbuf->paddr);
2315         rxd->wb.upper.status_error = 0;
2316 }
2317
2318 static int
2319 igb_newbuf(struct igb_rx_ring *rxr, int i, boolean_t wait)
2320 {
2321         struct mbuf *m;
2322         bus_dma_segment_t seg;
2323         bus_dmamap_t map;
2324         struct igb_rx_buf *rxbuf;
2325         int error, nseg;
2326
2327         m = m_getcl(wait ? M_WAITOK : M_NOWAIT, MT_DATA, M_PKTHDR);
2328         if (m == NULL) {
2329                 if (wait) {
2330                         if_printf(&rxr->sc->arpcom.ac_if,
2331                             "Unable to allocate RX mbuf\n");
2332                 }
2333                 return ENOBUFS;
2334         }
2335         m->m_len = m->m_pkthdr.len = MCLBYTES;
2336
2337         if (rxr->sc->max_frame_size <= MCLBYTES - ETHER_ALIGN)
2338                 m_adj(m, ETHER_ALIGN);
2339
2340         error = bus_dmamap_load_mbuf_segment(rxr->rx_tag,
2341             rxr->rx_sparemap, m, &seg, 1, &nseg, BUS_DMA_NOWAIT);
2342         if (error) {
2343                 m_freem(m);
2344                 if (wait) {
2345                         if_printf(&rxr->sc->arpcom.ac_if,
2346                             "Unable to load RX mbuf\n");
2347                 }
2348                 return error;
2349         }
2350
2351         rxbuf = &rxr->rx_buf[i];
2352         if (rxbuf->m_head != NULL)
2353                 bus_dmamap_unload(rxr->rx_tag, rxbuf->map);
2354
2355         map = rxbuf->map;
2356         rxbuf->map = rxr->rx_sparemap;
2357         rxr->rx_sparemap = map;
2358
2359         rxbuf->m_head = m;
2360         rxbuf->paddr = seg.ds_addr;
2361
2362         igb_setup_rxdesc(&rxr->rx_base[i], rxbuf);
2363         return 0;
2364 }
2365
2366 static int
2367 igb_init_rx_ring(struct igb_rx_ring *rxr)
2368 {
2369         int i;
2370
2371         /* Clear the ring contents */
2372         bzero(rxr->rx_base,
2373             rxr->num_rx_desc * sizeof(union e1000_adv_rx_desc));
2374
2375         /* Now replenish the ring mbufs */
2376         for (i = 0; i < rxr->num_rx_desc; ++i) {
2377                 int error;
2378
2379                 error = igb_newbuf(rxr, i, TRUE);
2380                 if (error)
2381                         return error;
2382         }
2383
2384         /* Setup our descriptor indices */
2385         rxr->next_to_check = 0;
2386
2387         rxr->fmp = NULL;
2388         rxr->lmp = NULL;
2389         rxr->discard = FALSE;
2390
2391         return 0;
2392 }
2393
2394 static void
2395 igb_init_rx_unit(struct igb_softc *sc)
2396 {
2397         struct ifnet *ifp = &sc->arpcom.ac_if;
2398         struct e1000_hw *hw = &sc->hw;
2399         uint32_t rctl, rxcsum, srrctl = 0;
2400         int i;
2401
2402         /*
2403          * Make sure receives are disabled while setting
2404          * up the descriptor ring
2405          */
2406         rctl = E1000_READ_REG(hw, E1000_RCTL);
2407         E1000_WRITE_REG(hw, E1000_RCTL, rctl & ~E1000_RCTL_EN);
2408
2409 #if 0
2410         /*
2411         ** Set up for header split
2412         */
2413         if (igb_header_split) {
2414                 /* Use a standard mbuf for the header */
2415                 srrctl |= IGB_HDR_BUF << E1000_SRRCTL_BSIZEHDRSIZE_SHIFT;
2416                 srrctl |= E1000_SRRCTL_DESCTYPE_HDR_SPLIT_ALWAYS;
2417         } else
2418 #endif
2419                 srrctl |= E1000_SRRCTL_DESCTYPE_ADV_ONEBUF;
2420
2421         /*
2422         ** Set up for jumbo frames
2423         */
2424         if (ifp->if_mtu > ETHERMTU) {
2425                 rctl |= E1000_RCTL_LPE;
2426 #if 0
2427                 if (adapter->rx_mbuf_sz == MJUMPAGESIZE) {
2428                         srrctl |= 4096 >> E1000_SRRCTL_BSIZEPKT_SHIFT;
2429                         rctl |= E1000_RCTL_SZ_4096 | E1000_RCTL_BSEX;
2430                 } else if (adapter->rx_mbuf_sz > MJUMPAGESIZE) {
2431                         srrctl |= 8192 >> E1000_SRRCTL_BSIZEPKT_SHIFT;
2432                         rctl |= E1000_RCTL_SZ_8192 | E1000_RCTL_BSEX;
2433                 }
2434                 /* Set maximum packet len */
2435                 psize = adapter->max_frame_size;
2436                 /* are we on a vlan? */
2437                 if (adapter->ifp->if_vlantrunk != NULL)
2438                         psize += VLAN_TAG_SIZE;
2439                 E1000_WRITE_REG(&adapter->hw, E1000_RLPML, psize);
2440 #else
2441                 srrctl |= 2048 >> E1000_SRRCTL_BSIZEPKT_SHIFT;
2442                 rctl |= E1000_RCTL_SZ_2048;
2443 #endif
2444         } else {
2445                 rctl &= ~E1000_RCTL_LPE;
2446                 srrctl |= 2048 >> E1000_SRRCTL_BSIZEPKT_SHIFT;
2447                 rctl |= E1000_RCTL_SZ_2048;
2448         }
2449
2450         /* Setup the Base and Length of the Rx Descriptor Rings */
2451         for (i = 0; i < sc->rx_ring_inuse; ++i) {
2452                 struct igb_rx_ring *rxr = &sc->rx_rings[i];
2453                 uint64_t bus_addr = rxr->rxdma.dma_paddr;
2454                 uint32_t rxdctl;
2455
2456                 E1000_WRITE_REG(hw, E1000_RDLEN(i),
2457                     rxr->num_rx_desc * sizeof(struct e1000_rx_desc));
2458                 E1000_WRITE_REG(hw, E1000_RDBAH(i),
2459                     (uint32_t)(bus_addr >> 32));
2460                 E1000_WRITE_REG(hw, E1000_RDBAL(i),
2461                     (uint32_t)bus_addr);
2462                 E1000_WRITE_REG(hw, E1000_SRRCTL(i), srrctl);
2463                 /* Enable this Queue */
2464                 rxdctl = E1000_READ_REG(hw, E1000_RXDCTL(i));
2465                 rxdctl |= E1000_RXDCTL_QUEUE_ENABLE;
2466                 rxdctl &= 0xFFF00000;
2467                 rxdctl |= IGB_RX_PTHRESH;
2468                 rxdctl |= IGB_RX_HTHRESH << 8;
2469                 /*
2470                  * Don't set WTHRESH to a value above 1 on 82576, see:
2471                  * 82576 specification update errata #26
2472                  */
2473                 rxdctl |= IGB_RX_WTHRESH << 16;
2474                 E1000_WRITE_REG(hw, E1000_RXDCTL(i), rxdctl);
2475         }
2476
2477         rxcsum = E1000_READ_REG(&sc->hw, E1000_RXCSUM);
2478         rxcsum &= ~(E1000_RXCSUM_PCSS_MASK | E1000_RXCSUM_IPPCSE);
2479
2480         /*
2481          * Receive Checksum Offload for TCP and UDP
2482          *
2483          * Checksum offloading is also enabled if multiple receive
2484          * queue is to be supported, since we need it to figure out
2485          * fragments.
2486          */
2487         if ((ifp->if_capenable & IFCAP_RXCSUM) || IGB_ENABLE_HWRSS(sc)) {
2488                 /*
2489                  * NOTE:
2490                  * PCSD must be enabled to enable multiple
2491                  * receive queues.
2492                  */
2493                 rxcsum |= E1000_RXCSUM_IPOFL | E1000_RXCSUM_TUOFL |
2494                     E1000_RXCSUM_PCSD;
2495         } else {
2496                 rxcsum &= ~(E1000_RXCSUM_IPOFL | E1000_RXCSUM_TUOFL |
2497                     E1000_RXCSUM_PCSD);
2498         }
2499         E1000_WRITE_REG(&sc->hw, E1000_RXCSUM, rxcsum);
2500
2501         if (IGB_ENABLE_HWRSS(sc)) {
2502                 uint8_t key[IGB_NRSSRK * IGB_RSSRK_SIZE];
2503                 uint32_t reta_shift;
2504                 int j, r;
2505
2506                 /*
2507                  * NOTE:
2508                  * When we reach here, RSS has already been disabled
2509                  * in igb_stop(), so we could safely configure RSS key
2510                  * and redirect table.
2511                  */
2512
2513                 /*
2514                  * Configure RSS key
2515                  */
2516                 toeplitz_get_key(key, sizeof(key));
2517                 for (i = 0; i < IGB_NRSSRK; ++i) {
2518                         uint32_t rssrk;
2519
2520                         rssrk = IGB_RSSRK_VAL(key, i);
2521                         IGB_RSS_DPRINTF(sc, 1, "rssrk%d 0x%08x\n", i, rssrk);
2522
2523                         E1000_WRITE_REG(hw, E1000_RSSRK(i), rssrk);
2524                 }
2525
2526                 /*
2527                  * Configure RSS redirect table in following fashion:
2528                  * (hash & ring_cnt_mask) == rdr_table[(hash & rdr_table_mask)]
2529                  */
2530                 reta_shift = IGB_RETA_SHIFT;
2531                 if (hw->mac.type == e1000_82575)
2532                         reta_shift = IGB_RETA_SHIFT_82575;
2533
2534                 r = 0;
2535                 for (j = 0; j < IGB_NRETA; ++j) {
2536                         uint32_t reta = 0;
2537
2538                         for (i = 0; i < IGB_RETA_SIZE; ++i) {
2539                                 uint32_t q;
2540
2541                                 q = (r % sc->rx_ring_inuse) << reta_shift;
2542                                 reta |= q << (8 * i);
2543                                 ++r;
2544                         }
2545                         IGB_RSS_DPRINTF(sc, 1, "reta 0x%08x\n", reta);
2546                         E1000_WRITE_REG(hw, E1000_RETA(j), reta);
2547                 }
2548
2549                 /*
2550                  * Enable multiple receive queues.
2551                  * Enable IPv4 RSS standard hash functions.
2552                  * Disable RSS interrupt on 82575
2553                  */
2554                 E1000_WRITE_REG(&sc->hw, E1000_MRQC,
2555                                 E1000_MRQC_ENABLE_RSS_4Q |
2556                                 E1000_MRQC_RSS_FIELD_IPV4_TCP |
2557                                 E1000_MRQC_RSS_FIELD_IPV4);
2558         }
2559
2560         /* Setup the Receive Control Register */
2561         rctl &= ~(3 << E1000_RCTL_MO_SHIFT);
2562         rctl |= E1000_RCTL_EN | E1000_RCTL_BAM | E1000_RCTL_LBM_NO |
2563             E1000_RCTL_RDMTS_HALF |
2564             (hw->mac.mc_filter_type << E1000_RCTL_MO_SHIFT);
2565         /* Strip CRC bytes. */
2566         rctl |= E1000_RCTL_SECRC;
2567         /* Make sure VLAN Filters are off */
2568         rctl &= ~E1000_RCTL_VFE;
2569         /* Don't store bad packets */
2570         rctl &= ~E1000_RCTL_SBP;
2571
2572         /* Enable Receives */
2573         E1000_WRITE_REG(hw, E1000_RCTL, rctl);
2574
2575         /*
2576          * Setup the HW Rx Head and Tail Descriptor Pointers
2577          *   - needs to be after enable
2578          */
2579         for (i = 0; i < sc->rx_ring_inuse; ++i) {
2580                 struct igb_rx_ring *rxr = &sc->rx_rings[i];
2581
2582                 E1000_WRITE_REG(hw, E1000_RDH(i), rxr->next_to_check);
2583                 E1000_WRITE_REG(hw, E1000_RDT(i), rxr->num_rx_desc - 1);
2584         }
2585 }
2586
2587 static void
2588 igb_rx_refresh(struct igb_rx_ring *rxr, int i)
2589 {
2590         if (--i < 0)
2591                 i = rxr->num_rx_desc - 1;
2592         E1000_WRITE_REG(&rxr->sc->hw, E1000_RDT(rxr->me), i);
2593 }
2594
2595 static void
2596 igb_rxeof(struct igb_rx_ring *rxr, int count)
2597 {
2598         struct ifnet *ifp = &rxr->sc->arpcom.ac_if;
2599         union e1000_adv_rx_desc *cur;
2600         uint32_t staterr;
2601         int i, ncoll = 0, cpuid = mycpuid;
2602
2603         i = rxr->next_to_check;
2604         cur = &rxr->rx_base[i];
2605         staterr = le32toh(cur->wb.upper.status_error);
2606
2607         if ((staterr & E1000_RXD_STAT_DD) == 0)
2608                 return;
2609
2610         while ((staterr & E1000_RXD_STAT_DD) && count != 0) {
2611                 struct pktinfo *pi = NULL, pi0;
2612                 struct igb_rx_buf *rxbuf = &rxr->rx_buf[i];
2613                 struct mbuf *m = NULL;
2614                 boolean_t eop;
2615
2616                 eop = (staterr & E1000_RXD_STAT_EOP) ? TRUE : FALSE;
2617                 if (eop)
2618                         --count;
2619
2620                 ++ncoll;
2621                 if ((staterr & E1000_RXDEXT_ERR_FRAME_ERR_MASK) == 0 &&
2622                     !rxr->discard) {
2623                         struct mbuf *mp = rxbuf->m_head;
2624                         uint32_t hash, hashtype;
2625                         uint16_t vlan;
2626                         int len;
2627
2628                         len = le16toh(cur->wb.upper.length);
2629                         if ((rxr->sc->hw.mac.type == e1000_i350 ||
2630                              rxr->sc->hw.mac.type == e1000_i354) &&
2631                             (staterr & E1000_RXDEXT_STATERR_LB))
2632                                 vlan = be16toh(cur->wb.upper.vlan);
2633                         else
2634                                 vlan = le16toh(cur->wb.upper.vlan);
2635
2636                         hash = le32toh(cur->wb.lower.hi_dword.rss);
2637                         hashtype = le32toh(cur->wb.lower.lo_dword.data) &
2638                             E1000_RXDADV_RSSTYPE_MASK;
2639
2640                         IGB_RSS_DPRINTF(rxr->sc, 10,
2641                             "ring%d, hash 0x%08x, hashtype %u\n",
2642                             rxr->me, hash, hashtype);
2643
2644                         bus_dmamap_sync(rxr->rx_tag, rxbuf->map,
2645                             BUS_DMASYNC_POSTREAD);
2646
2647                         if (igb_newbuf(rxr, i, FALSE) != 0) {
2648                                 IFNET_STAT_INC(ifp, iqdrops, 1);
2649                                 goto discard;
2650                         }
2651
2652                         mp->m_len = len;
2653                         if (rxr->fmp == NULL) {
2654                                 mp->m_pkthdr.len = len;
2655                                 rxr->fmp = mp;
2656                                 rxr->lmp = mp;
2657                         } else {
2658                                 rxr->lmp->m_next = mp;
2659                                 rxr->lmp = rxr->lmp->m_next;
2660                                 rxr->fmp->m_pkthdr.len += len;
2661                         }
2662
2663                         if (eop) {
2664                                 m = rxr->fmp;
2665                                 rxr->fmp = NULL;
2666                                 rxr->lmp = NULL;
2667
2668                                 m->m_pkthdr.rcvif = ifp;
2669                                 IFNET_STAT_INC(ifp, ipackets, 1);
2670
2671                                 if (ifp->if_capenable & IFCAP_RXCSUM)
2672                                         igb_rxcsum(staterr, m);
2673
2674                                 if (staterr & E1000_RXD_STAT_VP) {
2675                                         m->m_pkthdr.ether_vlantag = vlan;
2676                                         m->m_flags |= M_VLANTAG;
2677                                 }
2678
2679                                 if (ifp->if_capenable & IFCAP_RSS) {
2680                                         pi = igb_rssinfo(m, &pi0,
2681                                             hash, hashtype, staterr);
2682                                 }
2683 #ifdef IGB_RSS_DEBUG
2684                                 rxr->rx_packets++;
2685 #endif
2686                         }
2687                 } else {
2688                         IFNET_STAT_INC(ifp, ierrors, 1);
2689 discard:
2690                         igb_setup_rxdesc(cur, rxbuf);
2691                         if (!eop)
2692                                 rxr->discard = TRUE;
2693                         else
2694                                 rxr->discard = FALSE;
2695                         if (rxr->fmp != NULL) {
2696                                 m_freem(rxr->fmp);
2697                                 rxr->fmp = NULL;
2698                                 rxr->lmp = NULL;
2699                         }
2700                         m = NULL;
2701                 }
2702
2703                 if (m != NULL)
2704                         ifp->if_input(ifp, m, pi, cpuid);
2705
2706                 /* Advance our pointers to the next descriptor. */
2707                 if (++i == rxr->num_rx_desc)
2708                         i = 0;
2709
2710                 if (ncoll >= rxr->wreg_nsegs) {
2711                         igb_rx_refresh(rxr, i);
2712                         ncoll = 0;
2713                 }
2714
2715                 cur = &rxr->rx_base[i];
2716                 staterr = le32toh(cur->wb.upper.status_error);
2717         }
2718         rxr->next_to_check = i;
2719
2720         if (ncoll > 0)
2721                 igb_rx_refresh(rxr, i);
2722 }
2723
2724
2725 static void
2726 igb_set_vlan(struct igb_softc *sc)
2727 {
2728         struct e1000_hw *hw = &sc->hw;
2729         uint32_t reg;
2730 #if 0
2731         struct ifnet *ifp = sc->arpcom.ac_if;
2732 #endif
2733
2734         if (sc->vf_ifp) {
2735                 e1000_rlpml_set_vf(hw, sc->max_frame_size + VLAN_TAG_SIZE);
2736                 return;
2737         }
2738
2739         reg = E1000_READ_REG(hw, E1000_CTRL);
2740         reg |= E1000_CTRL_VME;
2741         E1000_WRITE_REG(hw, E1000_CTRL, reg);
2742
2743 #if 0
2744         /* Enable the Filter Table */
2745         if (ifp->if_capenable & IFCAP_VLAN_HWFILTER) {
2746                 reg = E1000_READ_REG(hw, E1000_RCTL);
2747                 reg &= ~E1000_RCTL_CFIEN;
2748                 reg |= E1000_RCTL_VFE;
2749                 E1000_WRITE_REG(hw, E1000_RCTL, reg);
2750         }
2751 #endif
2752
2753         /* Update the frame size */
2754         E1000_WRITE_REG(&sc->hw, E1000_RLPML,
2755             sc->max_frame_size + VLAN_TAG_SIZE);
2756
2757 #if 0
2758         /* Don't bother with table if no vlans */
2759         if ((adapter->num_vlans == 0) ||
2760             ((ifp->if_capenable & IFCAP_VLAN_HWFILTER) == 0))
2761                 return;
2762         /*
2763         ** A soft reset zero's out the VFTA, so
2764         ** we need to repopulate it now.
2765         */
2766         for (int i = 0; i < IGB_VFTA_SIZE; i++)
2767                 if (adapter->shadow_vfta[i] != 0) {
2768                         if (adapter->vf_ifp)
2769                                 e1000_vfta_set_vf(hw,
2770                                     adapter->shadow_vfta[i], TRUE);
2771                         else
2772                                 E1000_WRITE_REG_ARRAY(hw, E1000_VFTA,
2773                                  i, adapter->shadow_vfta[i]);
2774                 }
2775 #endif
2776 }
2777
2778 static void
2779 igb_enable_intr(struct igb_softc *sc)
2780 {
2781         if (sc->intr_type != PCI_INTR_TYPE_MSIX) {
2782                 lwkt_serialize_handler_enable(&sc->main_serialize);
2783         } else {
2784                 int i;
2785
2786                 for (i = 0; i < sc->msix_cnt; ++i) {
2787                         lwkt_serialize_handler_enable(
2788                             sc->msix_data[i].msix_serialize);
2789                 }
2790         }
2791
2792         if ((sc->flags & IGB_FLAG_SHARED_INTR) == 0) {
2793                 if (sc->intr_type == PCI_INTR_TYPE_MSIX)
2794                         E1000_WRITE_REG(&sc->hw, E1000_EIAC, sc->intr_mask);
2795                 else
2796                         E1000_WRITE_REG(&sc->hw, E1000_EIAC, 0);
2797                 E1000_WRITE_REG(&sc->hw, E1000_EIAM, sc->intr_mask);
2798                 E1000_WRITE_REG(&sc->hw, E1000_EIMS, sc->intr_mask);
2799                 E1000_WRITE_REG(&sc->hw, E1000_IMS, E1000_IMS_LSC);
2800         } else {
2801                 E1000_WRITE_REG(&sc->hw, E1000_IMS, IMS_ENABLE_MASK);
2802         }
2803         E1000_WRITE_FLUSH(&sc->hw);
2804 }
2805
2806 static void
2807 igb_disable_intr(struct igb_softc *sc)
2808 {
2809         if ((sc->flags & IGB_FLAG_SHARED_INTR) == 0) {
2810                 E1000_WRITE_REG(&sc->hw, E1000_EIMC, 0xffffffff);
2811                 E1000_WRITE_REG(&sc->hw, E1000_EIAC, 0);
2812         }
2813         E1000_WRITE_REG(&sc->hw, E1000_IMC, 0xffffffff);
2814         E1000_WRITE_FLUSH(&sc->hw);
2815
2816         if (sc->intr_type != PCI_INTR_TYPE_MSIX) {
2817                 lwkt_serialize_handler_disable(&sc->main_serialize);
2818         } else {
2819                 int i;
2820
2821                 for (i = 0; i < sc->msix_cnt; ++i) {
2822                         lwkt_serialize_handler_disable(
2823                             sc->msix_data[i].msix_serialize);
2824                 }
2825         }
2826 }
2827
2828 /*
2829  * Bit of a misnomer, what this really means is
2830  * to enable OS management of the system... aka
2831  * to disable special hardware management features
2832  */
2833 static void
2834 igb_get_mgmt(struct igb_softc *sc)
2835 {
2836         if (sc->flags & IGB_FLAG_HAS_MGMT) {
2837                 int manc2h = E1000_READ_REG(&sc->hw, E1000_MANC2H);
2838                 int manc = E1000_READ_REG(&sc->hw, E1000_MANC);
2839
2840                 /* disable hardware interception of ARP */
2841                 manc &= ~E1000_MANC_ARP_EN;
2842
2843                 /* enable receiving management packets to the host */
2844                 manc |= E1000_MANC_EN_MNG2HOST;
2845                 manc2h |= 1 << 5; /* Mng Port 623 */
2846                 manc2h |= 1 << 6; /* Mng Port 664 */
2847                 E1000_WRITE_REG(&sc->hw, E1000_MANC2H, manc2h);
2848                 E1000_WRITE_REG(&sc->hw, E1000_MANC, manc);
2849         }
2850 }
2851
2852 /*
2853  * Give control back to hardware management controller
2854  * if there is one.
2855  */
2856 static void
2857 igb_rel_mgmt(struct igb_softc *sc)
2858 {
2859         if (sc->flags & IGB_FLAG_HAS_MGMT) {
2860                 int manc = E1000_READ_REG(&sc->hw, E1000_MANC);
2861
2862                 /* Re-enable hardware interception of ARP */
2863                 manc |= E1000_MANC_ARP_EN;
2864                 manc &= ~E1000_MANC_EN_MNG2HOST;
2865
2866                 E1000_WRITE_REG(&sc->hw, E1000_MANC, manc);
2867         }
2868 }
2869
2870 /*
2871  * Sets CTRL_EXT:DRV_LOAD bit.
2872  *
2873  * For ASF and Pass Through versions of f/w this means that
2874  * the driver is loaded.
2875  */
2876 static void
2877 igb_get_hw_control(struct igb_softc *sc)
2878 {
2879         uint32_t ctrl_ext;
2880
2881         if (sc->vf_ifp)
2882                 return;
2883
2884         /* Let firmware know the driver has taken over */
2885         ctrl_ext = E1000_READ_REG(&sc->hw, E1000_CTRL_EXT);
2886         E1000_WRITE_REG(&sc->hw, E1000_CTRL_EXT,
2887             ctrl_ext | E1000_CTRL_EXT_DRV_LOAD);
2888 }
2889
2890 /*
2891  * Resets CTRL_EXT:DRV_LOAD bit.
2892  *
2893  * For ASF and Pass Through versions of f/w this means that the
2894  * driver is no longer loaded.
2895  */
2896 static void
2897 igb_rel_hw_control(struct igb_softc *sc)
2898 {
2899         uint32_t ctrl_ext;
2900
2901         if (sc->vf_ifp)
2902                 return;
2903
2904         /* Let firmware taken over control of h/w */
2905         ctrl_ext = E1000_READ_REG(&sc->hw, E1000_CTRL_EXT);
2906         E1000_WRITE_REG(&sc->hw, E1000_CTRL_EXT,
2907             ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD);
2908 }
2909
2910 static boolean_t
2911 igb_is_valid_ether_addr(const uint8_t *addr)
2912 {
2913         uint8_t zero_addr[ETHER_ADDR_LEN] = { 0, 0, 0, 0, 0, 0 };
2914
2915         if ((addr[0] & 1) || !bcmp(addr, zero_addr, ETHER_ADDR_LEN))
2916                 return FALSE;
2917         return TRUE;
2918 }
2919
2920 /*
2921  * Enable PCI Wake On Lan capability
2922  */
2923 static void
2924 igb_enable_wol(device_t dev)
2925 {
2926         uint16_t cap, status;
2927         uint8_t id;
2928
2929         /* First find the capabilities pointer*/
2930         cap = pci_read_config(dev, PCIR_CAP_PTR, 2);
2931
2932         /* Read the PM Capabilities */
2933         id = pci_read_config(dev, cap, 1);
2934         if (id != PCIY_PMG)     /* Something wrong */
2935                 return;
2936
2937         /*
2938          * OK, we have the power capabilities,
2939          * so now get the status register
2940          */
2941         cap += PCIR_POWER_STATUS;
2942         status = pci_read_config(dev, cap, 2);
2943         status |= PCIM_PSTAT_PME | PCIM_PSTAT_PMEENABLE;
2944         pci_write_config(dev, cap, status, 2);
2945 }
2946
2947 static void
2948 igb_update_stats_counters(struct igb_softc *sc)
2949 {
2950         struct e1000_hw *hw = &sc->hw;
2951         struct e1000_hw_stats *stats;
2952         struct ifnet *ifp = &sc->arpcom.ac_if;
2953
2954         /*
2955          * The virtual function adapter has only a
2956          * small controlled set of stats, do only
2957          * those and return.
2958          */
2959         if (sc->vf_ifp) {
2960                 igb_update_vf_stats_counters(sc);
2961                 return;
2962         }
2963         stats = sc->stats;
2964
2965         if (sc->hw.phy.media_type == e1000_media_type_copper ||
2966             (E1000_READ_REG(hw, E1000_STATUS) & E1000_STATUS_LU)) {
2967                 stats->symerrs +=
2968                     E1000_READ_REG(hw,E1000_SYMERRS);
2969                 stats->sec += E1000_READ_REG(hw, E1000_SEC);
2970         }
2971
2972         stats->crcerrs += E1000_READ_REG(hw, E1000_CRCERRS);
2973         stats->mpc += E1000_READ_REG(hw, E1000_MPC);
2974         stats->scc += E1000_READ_REG(hw, E1000_SCC);
2975         stats->ecol += E1000_READ_REG(hw, E1000_ECOL);
2976
2977         stats->mcc += E1000_READ_REG(hw, E1000_MCC);
2978         stats->latecol += E1000_READ_REG(hw, E1000_LATECOL);
2979         stats->colc += E1000_READ_REG(hw, E1000_COLC);
2980         stats->dc += E1000_READ_REG(hw, E1000_DC);
2981         stats->rlec += E1000_READ_REG(hw, E1000_RLEC);
2982         stats->xonrxc += E1000_READ_REG(hw, E1000_XONRXC);
2983         stats->xontxc += E1000_READ_REG(hw, E1000_XONTXC);
2984
2985         /*
2986          * For watchdog management we need to know if we have been
2987          * paused during the last interval, so capture that here.
2988          */
2989         sc->pause_frames = E1000_READ_REG(hw, E1000_XOFFRXC);
2990         stats->xoffrxc += sc->pause_frames;
2991         stats->xofftxc += E1000_READ_REG(hw, E1000_XOFFTXC);
2992         stats->fcruc += E1000_READ_REG(hw, E1000_FCRUC);
2993         stats->prc64 += E1000_READ_REG(hw, E1000_PRC64);
2994         stats->prc127 += E1000_READ_REG(hw, E1000_PRC127);
2995         stats->prc255 += E1000_READ_REG(hw, E1000_PRC255);
2996         stats->prc511 += E1000_READ_REG(hw, E1000_PRC511);
2997         stats->prc1023 += E1000_READ_REG(hw, E1000_PRC1023);
2998         stats->prc1522 += E1000_READ_REG(hw, E1000_PRC1522);
2999         stats->gprc += E1000_READ_REG(hw, E1000_GPRC);
3000         stats->bprc += E1000_READ_REG(hw, E1000_BPRC);
3001         stats->mprc += E1000_READ_REG(hw, E1000_MPRC);
3002         stats->gptc += E1000_READ_REG(hw, E1000_GPTC);
3003
3004         /* For the 64-bit byte counters the low dword must be read first. */
3005         /* Both registers clear on the read of the high dword */
3006
3007         stats->gorc += E1000_READ_REG(hw, E1000_GORCL) +
3008             ((uint64_t)E1000_READ_REG(hw, E1000_GORCH) << 32);
3009         stats->gotc += E1000_READ_REG(hw, E1000_GOTCL) +
3010             ((uint64_t)E1000_READ_REG(hw, E1000_GOTCH) << 32);
3011
3012         stats->rnbc += E1000_READ_REG(hw, E1000_RNBC);
3013         stats->ruc += E1000_READ_REG(hw, E1000_RUC);
3014         stats->rfc += E1000_READ_REG(hw, E1000_RFC);
3015         stats->roc += E1000_READ_REG(hw, E1000_ROC);
3016         stats->rjc += E1000_READ_REG(hw, E1000_RJC);
3017
3018         stats->tor += E1000_READ_REG(hw, E1000_TORH);
3019         stats->tot += E1000_READ_REG(hw, E1000_TOTH);
3020
3021         stats->tpr += E1000_READ_REG(hw, E1000_TPR);
3022         stats->tpt += E1000_READ_REG(hw, E1000_TPT);
3023         stats->ptc64 += E1000_READ_REG(hw, E1000_PTC64);
3024         stats->ptc127 += E1000_READ_REG(hw, E1000_PTC127);
3025         stats->ptc255 += E1000_READ_REG(hw, E1000_PTC255);
3026         stats->ptc511 += E1000_READ_REG(hw, E1000_PTC511);
3027         stats->ptc1023 += E1000_READ_REG(hw, E1000_PTC1023);
3028         stats->ptc1522 += E1000_READ_REG(hw, E1000_PTC1522);
3029         stats->mptc += E1000_READ_REG(hw, E1000_MPTC);
3030         stats->bptc += E1000_READ_REG(hw, E1000_BPTC);
3031
3032         /* Interrupt Counts */
3033
3034         stats->iac += E1000_READ_REG(hw, E1000_IAC);
3035         stats->icrxptc += E1000_READ_REG(hw, E1000_ICRXPTC);
3036         stats->icrxatc += E1000_READ_REG(hw, E1000_ICRXATC);
3037         stats->ictxptc += E1000_READ_REG(hw, E1000_ICTXPTC);
3038         stats->ictxatc += E1000_READ_REG(hw, E1000_ICTXATC);
3039         stats->ictxqec += E1000_READ_REG(hw, E1000_ICTXQEC);
3040         stats->ictxqmtc += E1000_READ_REG(hw, E1000_ICTXQMTC);
3041         stats->icrxdmtc += E1000_READ_REG(hw, E1000_ICRXDMTC);
3042         stats->icrxoc += E1000_READ_REG(hw, E1000_ICRXOC);
3043
3044         /* Host to Card Statistics */
3045
3046         stats->cbtmpc += E1000_READ_REG(hw, E1000_CBTMPC);
3047         stats->htdpmc += E1000_READ_REG(hw, E1000_HTDPMC);
3048         stats->cbrdpc += E1000_READ_REG(hw, E1000_CBRDPC);
3049         stats->cbrmpc += E1000_READ_REG(hw, E1000_CBRMPC);
3050         stats->rpthc += E1000_READ_REG(hw, E1000_RPTHC);
3051         stats->hgptc += E1000_READ_REG(hw, E1000_HGPTC);
3052         stats->htcbdpc += E1000_READ_REG(hw, E1000_HTCBDPC);
3053         stats->hgorc += (E1000_READ_REG(hw, E1000_HGORCL) +
3054             ((uint64_t)E1000_READ_REG(hw, E1000_HGORCH) << 32));
3055         stats->hgotc += (E1000_READ_REG(hw, E1000_HGOTCL) +
3056             ((uint64_t)E1000_READ_REG(hw, E1000_HGOTCH) << 32));
3057         stats->lenerrs += E1000_READ_REG(hw, E1000_LENERRS);
3058         stats->scvpc += E1000_READ_REG(hw, E1000_SCVPC);
3059         stats->hrmpc += E1000_READ_REG(hw, E1000_HRMPC);
3060
3061         stats->algnerrc += E1000_READ_REG(hw, E1000_ALGNERRC);
3062         stats->rxerrc += E1000_READ_REG(hw, E1000_RXERRC);
3063         stats->tncrs += E1000_READ_REG(hw, E1000_TNCRS);
3064         stats->cexterr += E1000_READ_REG(hw, E1000_CEXTERR);
3065         stats->tsctc += E1000_READ_REG(hw, E1000_TSCTC);
3066         stats->tsctfc += E1000_READ_REG(hw, E1000_TSCTFC);
3067
3068         IFNET_STAT_SET(ifp, collisions, stats->colc);
3069
3070         /* Rx Errors */
3071         IFNET_STAT_SET(ifp, ierrors,
3072             stats->rxerrc + stats->crcerrs + stats->algnerrc +
3073             stats->ruc + stats->roc + stats->mpc + stats->cexterr);
3074
3075         /* Tx Errors */
3076         IFNET_STAT_SET(ifp, oerrors,
3077             stats->ecol + stats->latecol + sc->watchdog_events);
3078
3079         /* Driver specific counters */
3080         sc->device_control = E1000_READ_REG(hw, E1000_CTRL);
3081         sc->rx_control = E1000_READ_REG(hw, E1000_RCTL);
3082         sc->int_mask = E1000_READ_REG(hw, E1000_IMS);
3083         sc->eint_mask = E1000_READ_REG(hw, E1000_EIMS);
3084         sc->packet_buf_alloc_tx =
3085             ((E1000_READ_REG(hw, E1000_PBA) & 0xffff0000) >> 16);
3086         sc->packet_buf_alloc_rx =
3087             (E1000_READ_REG(hw, E1000_PBA) & 0xffff);
3088 }
3089
3090 static void
3091 igb_vf_init_stats(struct igb_softc *sc)
3092 {
3093         struct e1000_hw *hw = &sc->hw;
3094         struct e1000_vf_stats *stats;
3095
3096         stats = sc->stats;
3097         stats->last_gprc = E1000_READ_REG(hw, E1000_VFGPRC);
3098         stats->last_gorc = E1000_READ_REG(hw, E1000_VFGORC);
3099         stats->last_gptc = E1000_READ_REG(hw, E1000_VFGPTC);
3100         stats->last_gotc = E1000_READ_REG(hw, E1000_VFGOTC);
3101         stats->last_mprc = E1000_READ_REG(hw, E1000_VFMPRC);
3102 }
3103
3104 static void
3105 igb_update_vf_stats_counters(struct igb_softc *sc)
3106 {
3107         struct e1000_hw *hw = &sc->hw;
3108         struct e1000_vf_stats *stats;
3109
3110         if (sc->link_speed == 0)
3111                 return;
3112
3113         stats = sc->stats;
3114         UPDATE_VF_REG(E1000_VFGPRC, stats->last_gprc, stats->gprc);
3115         UPDATE_VF_REG(E1000_VFGORC, stats->last_gorc, stats->gorc);
3116         UPDATE_VF_REG(E1000_VFGPTC, stats->last_gptc, stats->gptc);
3117         UPDATE_VF_REG(E1000_VFGOTC, stats->last_gotc, stats->gotc);
3118         UPDATE_VF_REG(E1000_VFMPRC, stats->last_mprc, stats->mprc);
3119 }
3120
3121 #ifdef IFPOLL_ENABLE
3122
3123 static void
3124 igb_npoll_status(struct ifnet *ifp)
3125 {
3126         struct igb_softc *sc = ifp->if_softc;
3127         uint32_t reg_icr;
3128
3129         ASSERT_SERIALIZED(&sc->main_serialize);
3130
3131         reg_icr = E1000_READ_REG(&sc->hw, E1000_ICR);
3132         if (reg_icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC)) {
3133                 sc->hw.mac.get_link_status = 1;
3134                 igb_update_link_status(sc);
3135         }
3136 }
3137
3138 static void
3139 igb_npoll_tx(struct ifnet *ifp, void *arg, int cycle __unused)
3140 {
3141         struct igb_tx_ring *txr = arg;
3142
3143         ASSERT_SERIALIZED(&txr->tx_serialize);
3144
3145         igb_txeof(txr);
3146         if (!ifsq_is_empty(txr->ifsq))
3147                 ifsq_devstart(txr->ifsq);
3148 }
3149
3150 static void
3151 igb_npoll_rx(struct ifnet *ifp __unused, void *arg, int cycle)
3152 {
3153         struct igb_rx_ring *rxr = arg;
3154
3155         ASSERT_SERIALIZED(&rxr->rx_serialize);
3156
3157         igb_rxeof(rxr, cycle);
3158 }
3159
3160 static void
3161 igb_npoll(struct ifnet *ifp, struct ifpoll_info *info)
3162 {
3163         struct igb_softc *sc = ifp->if_softc;
3164         int i, txr_cnt, rxr_cnt;
3165
3166         ASSERT_IFNET_SERIALIZED_ALL(ifp);
3167
3168         if (info) {
3169                 int off;
3170
3171                 info->ifpi_status.status_func = igb_npoll_status;
3172                 info->ifpi_status.serializer = &sc->main_serialize;
3173
3174                 txr_cnt = igb_get_txring_inuse(sc, TRUE);
3175                 off = sc->tx_npoll_off;
3176                 for (i = 0; i < txr_cnt; ++i) {
3177                         struct igb_tx_ring *txr = &sc->tx_rings[i];
3178                         int idx = i + off;
3179
3180                         KKASSERT(idx < ncpus2);
3181                         info->ifpi_tx[idx].poll_func = igb_npoll_tx;
3182                         info->ifpi_tx[idx].arg = txr;
3183                         info->ifpi_tx[idx].serializer = &txr->tx_serialize;
3184                         ifsq_set_cpuid(txr->ifsq, idx);
3185                 }
3186
3187                 rxr_cnt = igb_get_rxring_inuse(sc, TRUE);
3188                 off = sc->rx_npoll_off;
3189                 for (i = 0; i < rxr_cnt; ++i) {
3190                         struct igb_rx_ring *rxr = &sc->rx_rings[i];
3191                         int idx = i + off;
3192
3193                         KKASSERT(idx < ncpus2);
3194                         info->ifpi_rx[idx].poll_func = igb_npoll_rx;
3195                         info->ifpi_rx[idx].arg = rxr;
3196                         info->ifpi_rx[idx].serializer = &rxr->rx_serialize;
3197                 }
3198
3199                 if (ifp->if_flags & IFF_RUNNING) {
3200                         if (rxr_cnt == sc->rx_ring_inuse &&
3201                             txr_cnt == sc->tx_ring_inuse) {
3202                                 igb_set_timer_cpuid(sc, TRUE);
3203                                 igb_disable_intr(sc);
3204                         } else {
3205                                 igb_init(sc);
3206                         }
3207                 }
3208         } else {
3209                 for (i = 0; i < sc->tx_ring_cnt; ++i) {
3210                         struct igb_tx_ring *txr = &sc->tx_rings[i];
3211
3212                         ifsq_set_cpuid(txr->ifsq, txr->tx_intr_cpuid);
3213                 }
3214
3215                 if (ifp->if_flags & IFF_RUNNING) {
3216                         txr_cnt = igb_get_txring_inuse(sc, FALSE);
3217                         rxr_cnt = igb_get_rxring_inuse(sc, FALSE);
3218
3219                         if (rxr_cnt == sc->rx_ring_inuse &&
3220                             txr_cnt == sc->tx_ring_inuse) {
3221                                 igb_set_timer_cpuid(sc, FALSE);
3222                                 igb_enable_intr(sc);
3223                         } else {
3224                                 igb_init(sc);
3225                         }
3226                 }
3227         }
3228 }
3229
3230 #endif /* IFPOLL_ENABLE */
3231
3232 static void
3233 igb_intr(void *xsc)
3234 {
3235         struct igb_softc *sc = xsc;
3236         struct ifnet *ifp = &sc->arpcom.ac_if;
3237         uint32_t eicr;
3238
3239         ASSERT_SERIALIZED(&sc->main_serialize);
3240
3241         eicr = E1000_READ_REG(&sc->hw, E1000_EICR);
3242
3243         if (eicr == 0)
3244                 return;
3245
3246         if (ifp->if_flags & IFF_RUNNING) {
3247                 struct igb_tx_ring *txr = &sc->tx_rings[0];
3248                 int i;
3249
3250                 for (i = 0; i < sc->rx_ring_inuse; ++i) {
3251                         struct igb_rx_ring *rxr = &sc->rx_rings[i];
3252
3253                         if (eicr & rxr->rx_intr_mask) {
3254                                 lwkt_serialize_enter(&rxr->rx_serialize);
3255                                 igb_rxeof(rxr, -1);
3256                                 lwkt_serialize_exit(&rxr->rx_serialize);
3257                         }
3258                 }
3259
3260                 if (eicr & txr->tx_intr_mask) {
3261                         lwkt_serialize_enter(&txr->tx_serialize);
3262                         igb_txeof(txr);
3263                         if (!ifsq_is_empty(txr->ifsq))
3264                                 ifsq_devstart(txr->ifsq);
3265                         lwkt_serialize_exit(&txr->tx_serialize);
3266                 }
3267         }
3268
3269         if (eicr & E1000_EICR_OTHER) {
3270                 uint32_t icr = E1000_READ_REG(&sc->hw, E1000_ICR);
3271
3272                 /* Link status change */
3273                 if (icr & E1000_ICR_LSC) {
3274                         sc->hw.mac.get_link_status = 1;
3275                         igb_update_link_status(sc);
3276                 }
3277         }
3278
3279         /*
3280          * Reading EICR has the side effect to clear interrupt mask,
3281          * so all interrupts need to be enabled here.
3282          */
3283         E1000_WRITE_REG(&sc->hw, E1000_EIMS, sc->intr_mask);
3284 }
3285
3286 static void
3287 igb_intr_shared(void *xsc)
3288 {
3289         struct igb_softc *sc = xsc;
3290         struct ifnet *ifp = &sc->arpcom.ac_if;
3291         uint32_t reg_icr;
3292
3293         ASSERT_SERIALIZED(&sc->main_serialize);
3294
3295         reg_icr = E1000_READ_REG(&sc->hw, E1000_ICR);
3296
3297         /* Hot eject?  */
3298         if (reg_icr == 0xffffffff)
3299                 return;
3300
3301         /* Definitely not our interrupt.  */
3302         if (reg_icr == 0x0)
3303                 return;
3304
3305         if ((reg_icr & E1000_ICR_INT_ASSERTED) == 0)
3306                 return;
3307
3308         if (ifp->if_flags & IFF_RUNNING) {
3309                 if (reg_icr &
3310                     (E1000_ICR_RXT0 | E1000_ICR_RXDMT0 | E1000_ICR_RXO)) {
3311                         int i;
3312
3313                         for (i = 0; i < sc->rx_ring_inuse; ++i) {
3314                                 struct igb_rx_ring *rxr = &sc->rx_rings[i];
3315
3316                                 lwkt_serialize_enter(&rxr->rx_serialize);
3317                                 igb_rxeof(rxr, -1);
3318                                 lwkt_serialize_exit(&rxr->rx_serialize);
3319                         }
3320                 }
3321
3322                 if (reg_icr & E1000_ICR_TXDW) {
3323                         struct igb_tx_ring *txr = &sc->tx_rings[0];
3324
3325                         lwkt_serialize_enter(&txr->tx_serialize);
3326                         igb_txeof(txr);
3327                         if (!ifsq_is_empty(txr->ifsq))
3328                                 ifsq_devstart(txr->ifsq);
3329                         lwkt_serialize_exit(&txr->tx_serialize);
3330                 }
3331         }
3332
3333         /* Link status change */
3334         if (reg_icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC)) {
3335                 sc->hw.mac.get_link_status = 1;
3336                 igb_update_link_status(sc);
3337         }
3338
3339         if (reg_icr & E1000_ICR_RXO)
3340                 sc->rx_overruns++;
3341 }
3342
3343 static int
3344 igb_encap(struct igb_tx_ring *txr, struct mbuf **m_headp,
3345     int *segs_used, int *idx)
3346 {
3347         bus_dma_segment_t segs[IGB_MAX_SCATTER];
3348         bus_dmamap_t map;
3349         struct igb_tx_buf *tx_buf, *tx_buf_mapped;
3350         union e1000_adv_tx_desc *txd = NULL;
3351         struct mbuf *m_head = *m_headp;
3352         uint32_t olinfo_status = 0, cmd_type_len = 0, cmd_rs = 0;
3353         int maxsegs, nsegs, i, j, error;
3354         uint32_t hdrlen = 0;
3355
3356         if (m_head->m_pkthdr.csum_flags & CSUM_TSO) {
3357                 error = igb_tso_pullup(txr, m_headp);
3358                 if (error)
3359                         return error;
3360                 m_head = *m_headp;
3361         }
3362
3363         /* Set basic descriptor constants */
3364         cmd_type_len |= E1000_ADVTXD_DTYP_DATA;
3365         cmd_type_len |= E1000_ADVTXD_DCMD_IFCS | E1000_ADVTXD_DCMD_DEXT;
3366         if (m_head->m_flags & M_VLANTAG)
3367                 cmd_type_len |= E1000_ADVTXD_DCMD_VLE;
3368
3369         /*
3370          * Map the packet for DMA.
3371          */
3372         tx_buf = &txr->tx_buf[txr->next_avail_desc];
3373         tx_buf_mapped = tx_buf;
3374         map = tx_buf->map;
3375
3376         maxsegs = txr->tx_avail - IGB_TX_RESERVED;
3377         KASSERT(maxsegs >= txr->spare_desc, ("not enough spare TX desc\n"));
3378         if (maxsegs > IGB_MAX_SCATTER)
3379                 maxsegs = IGB_MAX_SCATTER;
3380
3381         error = bus_dmamap_load_mbuf_defrag(txr->tx_tag, map, m_headp,
3382             segs, maxsegs, &nsegs, BUS_DMA_NOWAIT);
3383         if (error) {
3384                 if (error == ENOBUFS)
3385                         txr->sc->mbuf_defrag_failed++;
3386                 else
3387                         txr->sc->no_tx_dma_setup++;
3388
3389                 m_freem(*m_headp);
3390                 *m_headp = NULL;
3391                 return error;
3392         }
3393         bus_dmamap_sync(txr->tx_tag, map, BUS_DMASYNC_PREWRITE);
3394
3395         m_head = *m_headp;
3396
3397         /*
3398          * Set up the TX context descriptor, if any hardware offloading is
3399          * needed.  This includes CSUM, VLAN, and TSO.  It will consume one
3400          * TX descriptor.
3401          *
3402          * Unlike these chips' predecessors (em/emx), TX context descriptor
3403          * will _not_ interfere TX data fetching pipelining.
3404          */
3405         if (m_head->m_pkthdr.csum_flags & CSUM_TSO) {
3406                 igb_tso_ctx(txr, m_head, &hdrlen);
3407                 cmd_type_len |= E1000_ADVTXD_DCMD_TSE;
3408                 olinfo_status |= E1000_TXD_POPTS_IXSM << 8;
3409                 olinfo_status |= E1000_TXD_POPTS_TXSM << 8;
3410                 txr->tx_nsegs++;
3411                 (*segs_used)++;
3412         } else if (igb_txcsum_ctx(txr, m_head)) {
3413                 if (m_head->m_pkthdr.csum_flags & CSUM_IP)
3414                         olinfo_status |= (E1000_TXD_POPTS_IXSM << 8);
3415                 if (m_head->m_pkthdr.csum_flags & (CSUM_UDP | CSUM_TCP))
3416                         olinfo_status |= (E1000_TXD_POPTS_TXSM << 8);
3417                 txr->tx_nsegs++;
3418                 (*segs_used)++;
3419         }
3420
3421         *segs_used += nsegs;
3422         txr->tx_nsegs += nsegs;
3423         if (txr->tx_nsegs >= txr->intr_nsegs) {
3424                 /*
3425                  * Report Status (RS) is turned on every intr_nsegs
3426                  * descriptors (roughly).
3427                  */
3428                 txr->tx_nsegs = 0;
3429                 cmd_rs = E1000_ADVTXD_DCMD_RS;
3430         }
3431
3432         /* Calculate payload length */
3433         olinfo_status |= ((m_head->m_pkthdr.len - hdrlen)
3434             << E1000_ADVTXD_PAYLEN_SHIFT);
3435
3436         /*
3437          * 82575 needs the TX context index added; the queue
3438          * index is used as TX context index here.
3439          */
3440         if (txr->sc->hw.mac.type == e1000_82575)
3441                 olinfo_status |= txr->me << 4;
3442
3443         /* Set up our transmit descriptors */
3444         i = txr->next_avail_desc;
3445         for (j = 0; j < nsegs; j++) {
3446                 bus_size_t seg_len;
3447                 bus_addr_t seg_addr;
3448
3449                 tx_buf = &txr->tx_buf[i];
3450                 txd = (union e1000_adv_tx_desc *)&txr->tx_base[i];
3451                 seg_addr = segs[j].ds_addr;
3452                 seg_len = segs[j].ds_len;
3453
3454                 txd->read.buffer_addr = htole64(seg_addr);
3455                 txd->read.cmd_type_len = htole32(cmd_type_len | seg_len);
3456                 txd->read.olinfo_status = htole32(olinfo_status);
3457                 if (++i == txr->num_tx_desc)
3458                         i = 0;
3459                 tx_buf->m_head = NULL;
3460         }
3461
3462         KASSERT(txr->tx_avail > nsegs, ("invalid avail TX desc\n"));
3463         txr->next_avail_desc = i;
3464         txr->tx_avail -= nsegs;
3465
3466         tx_buf->m_head = m_head;
3467         tx_buf_mapped->map = tx_buf->map;
3468         tx_buf->map = map;
3469
3470         /*
3471          * Last Descriptor of Packet needs End Of Packet (EOP)
3472          */
3473         txd->read.cmd_type_len |= htole32(E1000_ADVTXD_DCMD_EOP | cmd_rs);
3474
3475         /*
3476          * Defer TDT updating, until enough descrptors are setup
3477          */
3478         *idx = i;
3479 #ifdef IGB_TSS_DEBUG
3480         ++txr->tx_packets;
3481 #endif
3482
3483         return 0;
3484 }
3485
3486 static void
3487 igb_start(struct ifnet *ifp, struct ifaltq_subque *ifsq)
3488 {
3489         struct igb_softc *sc = ifp->if_softc;
3490         struct igb_tx_ring *txr = ifsq_get_priv(ifsq);
3491         struct mbuf *m_head;
3492         int idx = -1, nsegs = 0;
3493
3494         KKASSERT(txr->ifsq == ifsq);
3495         ASSERT_SERIALIZED(&txr->tx_serialize);
3496
3497         if ((ifp->if_flags & IFF_RUNNING) == 0 || ifsq_is_oactive(ifsq))
3498                 return;
3499
3500         if (!sc->link_active || (txr->tx_flags & IGB_TXFLAG_ENABLED) == 0) {
3501                 ifsq_purge(ifsq);
3502                 return;
3503         }
3504
3505         if (!IGB_IS_NOT_OACTIVE(txr))
3506                 igb_txeof(txr);
3507
3508         while (!ifsq_is_empty(ifsq)) {
3509                 if (IGB_IS_OACTIVE(txr)) {
3510                         ifsq_set_oactive(ifsq);
3511                         /* Set watchdog on */
3512                         txr->tx_watchdog.wd_timer = 5;
3513                         break;
3514                 }
3515
3516                 m_head = ifsq_dequeue(ifsq);
3517                 if (m_head == NULL)
3518                         break;
3519
3520                 if (igb_encap(txr, &m_head, &nsegs, &idx)) {
3521                         IFNET_STAT_INC(ifp, oerrors, 1);
3522                         continue;
3523                 }
3524
3525                 /*
3526                  * TX interrupt are aggressively aggregated, so increasing
3527                  * opackets at TX interrupt time will make the opackets
3528                  * statistics vastly inaccurate; we do the opackets increment
3529                  * now.
3530                  */
3531                 IFNET_STAT_INC(ifp, opackets, 1);
3532
3533                 if (nsegs >= txr->wreg_nsegs) {
3534                         E1000_WRITE_REG(&txr->sc->hw, E1000_TDT(txr->me), idx);
3535                         idx = -1;
3536                         nsegs = 0;
3537                 }
3538
3539                 /* Send a copy of the frame to the BPF listener */
3540                 ETHER_BPF_MTAP(ifp, m_head);
3541         }
3542         if (idx >= 0)
3543                 E1000_WRITE_REG(&txr->sc->hw, E1000_TDT(txr->me), idx);
3544 }
3545
3546 static void
3547 igb_watchdog(struct ifaltq_subque *ifsq)
3548 {
3549         struct igb_tx_ring *txr = ifsq_get_priv(ifsq);
3550         struct ifnet *ifp = ifsq_get_ifp(ifsq);
3551         struct igb_softc *sc = ifp->if_softc;
3552         int i;
3553
3554         KKASSERT(txr->ifsq == ifsq);
3555         ASSERT_IFNET_SERIALIZED_ALL(ifp);
3556
3557         /*
3558          * If flow control has paused us since last checking
3559          * it invalidates the watchdog timing, so dont run it.
3560          */
3561         if (sc->pause_frames) {
3562                 sc->pause_frames = 0;
3563                 txr->tx_watchdog.wd_timer = 5;
3564                 return;
3565         }
3566
3567         if_printf(ifp, "Watchdog timeout -- resetting\n");
3568         if_printf(ifp, "Queue(%d) tdh = %d, hw tdt = %d\n", txr->me,
3569             E1000_READ_REG(&sc->hw, E1000_TDH(txr->me)),
3570             E1000_READ_REG(&sc->hw, E1000_TDT(txr->me)));
3571         if_printf(ifp, "TX(%d) desc avail = %d, "
3572             "Next TX to Clean = %d\n",
3573             txr->me, txr->tx_avail, txr->next_to_clean);
3574
3575         IFNET_STAT_INC(ifp, oerrors, 1);
3576         sc->watchdog_events++;
3577
3578         igb_init(sc);
3579         for (i = 0; i < sc->tx_ring_inuse; ++i)
3580                 ifsq_devstart_sched(sc->tx_rings[i].ifsq);
3581 }
3582
3583 static void
3584 igb_set_eitr(struct igb_softc *sc, int idx, int rate)
3585 {
3586         uint32_t eitr = 0;
3587
3588         if (rate > 0) {
3589                 if (sc->hw.mac.type == e1000_82575) {
3590                         eitr = 1000000000 / 256 / rate;
3591                         /*
3592                          * NOTE:
3593                          * Document is wrong on the 2 bits left shift
3594                          */
3595                 } else {
3596                         eitr = 1000000 / rate;
3597                         eitr <<= IGB_EITR_INTVL_SHIFT;
3598                 }
3599
3600                 if (eitr == 0) {
3601                         /* Don't disable it */
3602                         eitr = 1 << IGB_EITR_INTVL_SHIFT;
3603                 } else if (eitr > IGB_EITR_INTVL_MASK) {
3604                         /* Don't allow it to be too large */
3605                         eitr = IGB_EITR_INTVL_MASK;
3606                 }
3607         }
3608         if (sc->hw.mac.type == e1000_82575)
3609                 eitr |= eitr << 16;
3610         else
3611                 eitr |= E1000_EITR_CNT_IGNR;
3612         E1000_WRITE_REG(&sc->hw, E1000_EITR(idx), eitr);
3613 }
3614
3615 static int
3616 igb_sysctl_intr_rate(SYSCTL_HANDLER_ARGS)
3617 {
3618         struct igb_softc *sc = (void *)arg1;
3619         struct ifnet *ifp = &sc->arpcom.ac_if;
3620         int error, intr_rate;
3621
3622         intr_rate = sc->intr_rate;
3623         error = sysctl_handle_int(oidp, &intr_rate, 0, req);
3624         if (error || req->newptr == NULL)
3625                 return error;
3626         if (intr_rate < 0)
3627                 return EINVAL;
3628
3629         ifnet_serialize_all(ifp);
3630
3631         sc->intr_rate = intr_rate;
3632         if (ifp->if_flags & IFF_RUNNING)
3633                 igb_set_eitr(sc, 0, sc->intr_rate);
3634
3635         if (bootverbose)
3636                 if_printf(ifp, "interrupt rate set to %d/sec\n", sc->intr_rate);
3637
3638         ifnet_deserialize_all(ifp);
3639
3640         return 0;
3641 }
3642
3643 static int
3644 igb_sysctl_msix_rate(SYSCTL_HANDLER_ARGS)
3645 {
3646         struct igb_msix_data *msix = (void *)arg1;
3647         struct igb_softc *sc = msix->msix_sc;
3648         struct ifnet *ifp = &sc->arpcom.ac_if;
3649         int error, msix_rate;
3650
3651         msix_rate = msix->msix_rate;
3652         error = sysctl_handle_int(oidp, &msix_rate, 0, req);
3653         if (error || req->newptr == NULL)
3654                 return error;
3655         if (msix_rate < 0)
3656                 return EINVAL;
3657
3658         lwkt_serialize_enter(msix->msix_serialize);
3659
3660         msix->msix_rate = msix_rate;
3661         if (ifp->if_flags & IFF_RUNNING)
3662                 igb_set_eitr(sc, msix->msix_vector, msix->msix_rate);
3663
3664         if (bootverbose) {
3665                 if_printf(ifp, "%s set to %d/sec\n", msix->msix_rate_desc,
3666                     msix->msix_rate);
3667         }
3668
3669         lwkt_serialize_exit(msix->msix_serialize);
3670
3671         return 0;
3672 }
3673
3674 static int
3675 igb_sysctl_tx_intr_nsegs(SYSCTL_HANDLER_ARGS)
3676 {
3677         struct igb_softc *sc = (void *)arg1;
3678         struct ifnet *ifp = &sc->arpcom.ac_if;
3679         struct igb_tx_ring *txr = &sc->tx_rings[0];
3680         int error, nsegs;
3681
3682         nsegs = txr->intr_nsegs;
3683         error = sysctl_handle_int(oidp, &nsegs, 0, req);
3684         if (error || req->newptr == NULL)
3685                 return error;
3686         if (nsegs <= 0)
3687                 return EINVAL;
3688
3689         ifnet_serialize_all(ifp);
3690
3691         if (nsegs >= txr->num_tx_desc - txr->oact_lo_desc ||
3692             nsegs >= txr->oact_hi_desc - IGB_MAX_SCATTER) {
3693                 error = EINVAL;
3694         } else {
3695                 int i;
3696
3697                 error = 0;
3698                 for (i = 0; i < sc->tx_ring_cnt; ++i)
3699                         sc->tx_rings[i].intr_nsegs = nsegs;
3700         }
3701
3702         ifnet_deserialize_all(ifp);
3703
3704         return error;
3705 }
3706
3707 static int
3708 igb_sysctl_rx_wreg_nsegs(SYSCTL_HANDLER_ARGS)
3709 {
3710         struct igb_softc *sc = (void *)arg1;
3711         struct ifnet *ifp = &sc->arpcom.ac_if;
3712         int error, nsegs, i;
3713
3714         nsegs = sc->rx_rings[0].wreg_nsegs;
3715         error = sysctl_handle_int(oidp, &nsegs, 0, req);
3716         if (error || req->newptr == NULL)
3717                 return error;
3718
3719         ifnet_serialize_all(ifp);
3720         for (i = 0; i < sc->rx_ring_cnt; ++i)
3721                 sc->rx_rings[i].wreg_nsegs =nsegs;
3722         ifnet_deserialize_all(ifp);
3723
3724         return 0;
3725 }
3726
3727 static int
3728 igb_sysctl_tx_wreg_nsegs(SYSCTL_HANDLER_ARGS)
3729 {
3730         struct igb_softc *sc = (void *)arg1;
3731         struct ifnet *ifp = &sc->arpcom.ac_if;
3732         int error, nsegs, i;
3733
3734         nsegs = sc->tx_rings[0].wreg_nsegs;
3735         error = sysctl_handle_int(oidp, &nsegs, 0, req);
3736         if (error || req->newptr == NULL)
3737                 return error;
3738
3739         ifnet_serialize_all(ifp);
3740         for (i = 0; i < sc->tx_ring_cnt; ++i)
3741                 sc->tx_rings[i].wreg_nsegs =nsegs;
3742         ifnet_deserialize_all(ifp);
3743
3744         return 0;
3745 }
3746
3747 #ifdef IFPOLL_ENABLE
3748
3749 static int
3750 igb_sysctl_npoll_rxoff(SYSCTL_HANDLER_ARGS)
3751 {
3752         struct igb_softc *sc = (void *)arg1;
3753         struct ifnet *ifp = &sc->arpcom.ac_if;
3754         int error, off;
3755
3756         off = sc->rx_npoll_off;
3757         error = sysctl_handle_int(oidp, &off, 0, req);
3758         if (error || req->newptr == NULL)
3759                 return error;
3760         if (off < 0)
3761                 return EINVAL;
3762
3763         ifnet_serialize_all(ifp);
3764         if (off >= ncpus2 || off % sc->rx_ring_cnt != 0) {
3765                 error = EINVAL;
3766         } else {
3767                 error = 0;
3768                 sc->rx_npoll_off = off;
3769         }
3770         ifnet_deserialize_all(ifp);
3771
3772         return error;
3773 }
3774
3775 static int
3776 igb_sysctl_npoll_txoff(SYSCTL_HANDLER_ARGS)
3777 {
3778         struct igb_softc *sc = (void *)arg1;
3779         struct ifnet *ifp = &sc->arpcom.ac_if;
3780         int error, off;
3781
3782         off = sc->tx_npoll_off;
3783         error = sysctl_handle_int(oidp, &off, 0, req);
3784         if (error || req->newptr == NULL)
3785                 return error;
3786         if (off < 0)
3787                 return EINVAL;
3788
3789         ifnet_serialize_all(ifp);
3790         if (off >= ncpus2 || off % sc->tx_ring_cnt != 0) {
3791                 error = EINVAL;
3792         } else {
3793                 error = 0;
3794                 sc->tx_npoll_off = off;
3795         }
3796         ifnet_deserialize_all(ifp);
3797
3798         return error;
3799 }
3800
3801 #endif  /* IFPOLL_ENABLE */
3802
3803 static void
3804 igb_init_intr(struct igb_softc *sc)
3805 {
3806         igb_set_intr_mask(sc);
3807
3808         if ((sc->flags & IGB_FLAG_SHARED_INTR) == 0)
3809                 igb_init_unshared_intr(sc);
3810
3811         if (sc->intr_type != PCI_INTR_TYPE_MSIX) {
3812                 igb_set_eitr(sc, 0, sc->intr_rate);
3813         } else {
3814                 int i;
3815
3816                 for (i = 0; i < sc->msix_cnt; ++i)
3817                         igb_set_eitr(sc, i, sc->msix_data[i].msix_rate);
3818         }
3819 }
3820
3821 static void
3822 igb_init_unshared_intr(struct igb_softc *sc)
3823 {
3824         struct e1000_hw *hw = &sc->hw;
3825         const struct igb_rx_ring *rxr;
3826         const struct igb_tx_ring *txr;
3827         uint32_t ivar, index;
3828         int i;
3829
3830         /*
3831          * Enable extended mode
3832          */
3833         if (sc->hw.mac.type != e1000_82575) {
3834                 uint32_t gpie;
3835                 int ivar_max;
3836
3837                 gpie = E1000_GPIE_NSICR;
3838                 if (sc->intr_type == PCI_INTR_TYPE_MSIX) {
3839                         gpie |= E1000_GPIE_MSIX_MODE |
3840                             E1000_GPIE_EIAME |
3841                             E1000_GPIE_PBA;
3842                 }
3843                 E1000_WRITE_REG(hw, E1000_GPIE, gpie);
3844
3845                 /*
3846                  * Clear IVARs
3847                  */
3848                 switch (sc->hw.mac.type) {
3849                 case e1000_82576:
3850                         ivar_max = IGB_MAX_IVAR_82576;
3851                         break;
3852
3853                 case e1000_82580:
3854                         ivar_max = IGB_MAX_IVAR_82580;
3855                         break;
3856
3857                 case e1000_i350:
3858                         ivar_max = IGB_MAX_IVAR_I350;
3859                         break;
3860
3861                 case e1000_i354:
3862                         ivar_max = IGB_MAX_IVAR_I354;
3863                         break;
3864
3865                 case e1000_vfadapt:
3866                 case e1000_vfadapt_i350:
3867                         ivar_max = IGB_MAX_IVAR_VF;
3868                         break;
3869
3870                 case e1000_i210:
3871                         ivar_max = IGB_MAX_IVAR_I210;
3872                         break;
3873
3874                 case e1000_i211:
3875                         ivar_max = IGB_MAX_IVAR_I211;
3876                         break;
3877
3878                 default:
3879                         panic("unknown mac type %d\n", sc->hw.mac.type);
3880                 }
3881                 for (i = 0; i < ivar_max; ++i)
3882                         E1000_WRITE_REG_ARRAY(hw, E1000_IVAR0, i, 0);
3883                 E1000_WRITE_REG(hw, E1000_IVAR_MISC, 0);
3884         } else {
3885                 uint32_t tmp;
3886
3887                 KASSERT(sc->intr_type != PCI_INTR_TYPE_MSIX,
3888                     ("82575 w/ MSI-X"));
3889                 tmp = E1000_READ_REG(hw, E1000_CTRL_EXT);
3890                 tmp |= E1000_CTRL_EXT_IRCA;
3891                 E1000_WRITE_REG(hw, E1000_CTRL_EXT, tmp);
3892         }
3893
3894         /*
3895          * Map TX/RX interrupts to EICR
3896          */
3897         switch (sc->hw.mac.type) {
3898         case e1000_82580:
3899         case e1000_i350:
3900         case e1000_i354:
3901         case e1000_vfadapt:
3902         case e1000_vfadapt_i350:
3903         case e1000_i210:
3904         case e1000_i211:
3905                 /* RX entries */
3906                 for (i = 0; i < sc->rx_ring_inuse; ++i) {
3907                         rxr = &sc->rx_rings[i];
3908
3909                         index = i >> 1;
3910                         ivar = E1000_READ_REG_ARRAY(hw, E1000_IVAR0, index);
3911
3912                         if (i & 1) {
3913                                 ivar &= 0xff00ffff;
3914                                 ivar |=
3915                                 (rxr->rx_intr_bit | E1000_IVAR_VALID) << 16;
3916                         } else {
3917                                 ivar &= 0xffffff00;
3918                                 ivar |=
3919                                 (rxr->rx_intr_bit | E1000_IVAR_VALID);
3920                         }
3921                         E1000_WRITE_REG_ARRAY(hw, E1000_IVAR0, index, ivar);
3922                 }
3923                 /* TX entries */
3924                 for (i = 0; i < sc->tx_ring_inuse; ++i) {
3925                         txr = &sc->tx_rings[i];
3926
3927                         index = i >> 1;
3928                         ivar = E1000_READ_REG_ARRAY(hw, E1000_IVAR0, index);
3929
3930                         if (i & 1) {
3931                                 ivar &= 0x00ffffff;
3932                                 ivar |=
3933                                 (txr->tx_intr_bit | E1000_IVAR_VALID) << 24;
3934                         } else {
3935                                 ivar &= 0xffff00ff;
3936                                 ivar |=
3937                                 (txr->tx_intr_bit | E1000_IVAR_VALID) << 8;
3938                         }
3939                         E1000_WRITE_REG_ARRAY(hw, E1000_IVAR0, index, ivar);
3940                 }
3941                 if (sc->intr_type == PCI_INTR_TYPE_MSIX) {
3942                         ivar = (sc->sts_intr_bit | E1000_IVAR_VALID) << 8;
3943                         E1000_WRITE_REG(hw, E1000_IVAR_MISC, ivar);
3944                 }
3945                 break;
3946
3947         case e1000_82576:
3948                 /* RX entries */
3949                 for (i = 0; i < sc->rx_ring_inuse; ++i) {
3950                         rxr = &sc->rx_rings[i];
3951
3952                         index = i & 0x7; /* Each IVAR has two entries */
3953                         ivar = E1000_READ_REG_ARRAY(hw, E1000_IVAR0, index);
3954
3955                         if (i < 8) {
3956                                 ivar &= 0xffffff00;
3957                                 ivar |=
3958                                 (rxr->rx_intr_bit | E1000_IVAR_VALID);
3959                         } else {
3960                                 ivar &= 0xff00ffff;
3961                                 ivar |=
3962                                 (rxr->rx_intr_bit | E1000_IVAR_VALID) << 16;
3963                         }
3964                         E1000_WRITE_REG_ARRAY(hw, E1000_IVAR0, index, ivar);
3965                 }
3966                 /* TX entries */
3967                 for (i = 0; i < sc->tx_ring_inuse; ++i) {
3968                         txr = &sc->tx_rings[i];
3969
3970                         index = i & 0x7; /* Each IVAR has two entries */
3971                         ivar = E1000_READ_REG_ARRAY(hw, E1000_IVAR0, index);
3972
3973                         if (i < 8) {
3974                                 ivar &= 0xffff00ff;
3975                                 ivar |=
3976                                 (txr->tx_intr_bit | E1000_IVAR_VALID) << 8;
3977                         } else {
3978                                 ivar &= 0x00ffffff;
3979                                 ivar |=
3980                                 (txr->tx_intr_bit | E1000_IVAR_VALID) << 24;
3981                         }
3982                         E1000_WRITE_REG_ARRAY(hw, E1000_IVAR0, index, ivar);
3983                 }
3984                 if (sc->intr_type == PCI_INTR_TYPE_MSIX) {
3985                         ivar = (sc->sts_intr_bit | E1000_IVAR_VALID) << 8;
3986                         E1000_WRITE_REG(hw, E1000_IVAR_MISC, ivar);
3987                 }
3988                 break;
3989
3990         case e1000_82575:
3991                 /*
3992                  * Enable necessary interrupt bits.
3993                  *
3994                  * The name of the register is confusing; in addition to
3995                  * configuring the first vector of MSI-X, it also configures
3996                  * which bits of EICR could be set by the hardware even when
3997                  * MSI or line interrupt is used; it thus controls interrupt
3998                  * generation.  It MUST be configured explicitly; the default
3999                  * value mentioned in the datasheet is wrong: RX queue0 and
4000                  * TX queue0 are NOT enabled by default.
4001                  */
4002                 E1000_WRITE_REG(&sc->hw, E1000_MSIXBM(0), sc->intr_mask);
4003                 break;
4004
4005         default:
4006                 panic("unknown mac type %d\n", sc->hw.mac.type);
4007         }
4008 }
4009
4010 static int
4011 igb_setup_intr(struct igb_softc *sc)
4012 {
4013         int error;
4014
4015         if (sc->intr_type == PCI_INTR_TYPE_MSIX)
4016                 return igb_msix_setup(sc);
4017
4018         error = bus_setup_intr(sc->dev, sc->intr_res, INTR_MPSAFE,
4019             (sc->flags & IGB_FLAG_SHARED_INTR) ? igb_intr_shared : igb_intr,
4020             sc, &sc->intr_tag, &sc->main_serialize);
4021         if (error) {
4022                 device_printf(sc->dev, "Failed to register interrupt handler");
4023                 return error;
4024         }
4025         return 0;
4026 }
4027
4028 static void
4029 igb_set_txintr_mask(struct igb_tx_ring *txr, int *intr_bit0, int intr_bitmax)
4030 {
4031         if (txr->sc->hw.mac.type == e1000_82575) {
4032                 txr->tx_intr_bit = 0;   /* unused */
4033                 switch (txr->me) {
4034                 case 0:
4035                         txr->tx_intr_mask = E1000_EICR_TX_QUEUE0;
4036                         break;
4037                 case 1:
4038                         txr->tx_intr_mask = E1000_EICR_TX_QUEUE1;
4039                         break;
4040                 case 2:
4041                         txr->tx_intr_mask = E1000_EICR_TX_QUEUE2;
4042                         break;
4043                 case 3:
4044                         txr->tx_intr_mask = E1000_EICR_TX_QUEUE3;
4045                         break;
4046                 default:
4047                         panic("unsupported # of TX ring, %d\n", txr->me);
4048                 }
4049         } else {
4050                 int intr_bit = *intr_bit0;
4051
4052                 txr->tx_intr_bit = intr_bit % intr_bitmax;
4053                 txr->tx_intr_mask = 1 << txr->tx_intr_bit;
4054
4055                 *intr_bit0 = intr_bit + 1;
4056         }
4057 }
4058
4059 static void
4060 igb_set_rxintr_mask(struct igb_rx_ring *rxr, int *intr_bit0, int intr_bitmax)
4061 {
4062         if (rxr->sc->hw.mac.type == e1000_82575) {
4063                 rxr->rx_intr_bit = 0;   /* unused */
4064                 switch (rxr->me) {
4065                 case 0:
4066                         rxr->rx_intr_mask = E1000_EICR_RX_QUEUE0;
4067                         break;
4068                 case 1:
4069                         rxr->rx_intr_mask = E1000_EICR_RX_QUEUE1;
4070                         break;
4071                 case 2:
4072                         rxr->rx_intr_mask = E1000_EICR_RX_QUEUE2;
4073                         break;
4074                 case 3:
4075                         rxr->rx_intr_mask = E1000_EICR_RX_QUEUE3;
4076                         break;
4077                 default:
4078                         panic("unsupported # of RX ring, %d\n", rxr->me);
4079                 }
4080         } else {
4081                 int intr_bit = *intr_bit0;
4082
4083                 rxr->rx_intr_bit = intr_bit % intr_bitmax;
4084                 rxr->rx_intr_mask = 1 << rxr->rx_intr_bit;
4085
4086                 *intr_bit0 = intr_bit + 1;
4087         }
4088 }
4089
4090 static void
4091 igb_serialize(struct ifnet *ifp, enum ifnet_serialize slz)
4092 {
4093         struct igb_softc *sc = ifp->if_softc;
4094
4095         ifnet_serialize_array_enter(sc->serializes, sc->serialize_cnt, slz);
4096 }
4097
4098 static void
4099 igb_deserialize(struct ifnet *ifp, enum ifnet_serialize slz)
4100 {
4101         struct igb_softc *sc = ifp->if_softc;
4102
4103         ifnet_serialize_array_exit(sc->serializes, sc->serialize_cnt, slz);
4104 }
4105
4106 static int
4107 igb_tryserialize(struct ifnet *ifp, enum ifnet_serialize slz)
4108 {
4109         struct igb_softc *sc = ifp->if_softc;
4110
4111         return ifnet_serialize_array_try(sc->serializes, sc->serialize_cnt,
4112             slz);
4113 }
4114
4115 #ifdef INVARIANTS
4116
4117 static void
4118 igb_serialize_assert(struct ifnet *ifp, enum ifnet_serialize slz,
4119     boolean_t serialized)
4120 {
4121         struct igb_softc *sc = ifp->if_softc;
4122
4123         ifnet_serialize_array_assert(sc->serializes, sc->serialize_cnt,
4124             slz, serialized);
4125 }
4126
4127 #endif  /* INVARIANTS */
4128
4129 static void
4130 igb_set_intr_mask(struct igb_softc *sc)
4131 {
4132         int i;
4133
4134         sc->intr_mask = sc->sts_intr_mask;
4135         for (i = 0; i < sc->rx_ring_inuse; ++i)
4136                 sc->intr_mask |= sc->rx_rings[i].rx_intr_mask;
4137         for (i = 0; i < sc->tx_ring_inuse; ++i)
4138                 sc->intr_mask |= sc->tx_rings[i].tx_intr_mask;
4139         if (bootverbose) {
4140                 if_printf(&sc->arpcom.ac_if, "intr mask 0x%08x\n",
4141                     sc->intr_mask);
4142         }
4143 }
4144
4145 static int
4146 igb_alloc_intr(struct igb_softc *sc)
4147 {
4148         int i, intr_bit, intr_bitmax;
4149         u_int intr_flags;
4150
4151         igb_msix_try_alloc(sc);
4152         if (sc->intr_type == PCI_INTR_TYPE_MSIX)
4153                 goto done;
4154
4155         /*
4156          * Allocate MSI/legacy interrupt resource
4157          */
4158         sc->intr_type = pci_alloc_1intr(sc->dev, igb_msi_enable,
4159             &sc->intr_rid, &intr_flags);
4160
4161         if (sc->intr_type == PCI_INTR_TYPE_LEGACY) {
4162                 int unshared;
4163
4164                 unshared = device_getenv_int(sc->dev, "irq.unshared", 0);
4165                 if (!unshared) {
4166                         sc->flags |= IGB_FLAG_SHARED_INTR;
4167                         if (bootverbose)
4168                                 device_printf(sc->dev, "IRQ shared\n");
4169                 } else {
4170                         intr_flags &= ~RF_SHAREABLE;
4171                         if (bootverbose)
4172                                 device_printf(sc->dev, "IRQ unshared\n");
4173                 }
4174         }
4175
4176         sc->intr_res = bus_alloc_resource_any(sc->dev, SYS_RES_IRQ,
4177             &sc->intr_rid, intr_flags);
4178         if (sc->intr_res == NULL) {
4179                 device_printf(sc->dev, "Unable to allocate bus resource: "
4180                     "interrupt\n");
4181                 return ENXIO;
4182         }
4183
4184         for (i = 0; i < sc->tx_ring_cnt; ++i)
4185                 sc->tx_rings[i].tx_intr_cpuid = rman_get_cpuid(sc->intr_res);
4186
4187         /*
4188          * Setup MSI/legacy interrupt mask
4189          */
4190         switch (sc->hw.mac.type) {
4191         case e1000_82575:
4192                 intr_bitmax = IGB_MAX_TXRXINT_82575;
4193                 break;
4194
4195         case e1000_82576:
4196                 intr_bitmax = IGB_MAX_TXRXINT_82576;
4197                 break;
4198
4199         case e1000_82580:
4200                 intr_bitmax = IGB_MAX_TXRXINT_82580;
4201                 break;
4202
4203         case e1000_i350:
4204                 intr_bitmax = IGB_MAX_TXRXINT_I350;
4205                 break;
4206
4207         case e1000_i354:
4208                 intr_bitmax = IGB_MAX_TXRXINT_I354;
4209                 break;
4210
4211         case e1000_i210:
4212                 intr_bitmax = IGB_MAX_TXRXINT_I210;
4213                 break;
4214
4215         case e1000_i211:
4216                 intr_bitmax = IGB_MAX_TXRXINT_I211;
4217                 break;
4218
4219         default:
4220                 intr_bitmax = IGB_MIN_TXRXINT;
4221                 break;
4222         }
4223         intr_bit = 0;
4224         for (i = 0; i < sc->tx_ring_cnt; ++i)
4225                 igb_set_txintr_mask(&sc->tx_rings[i], &intr_bit, intr_bitmax);
4226         for (i = 0; i < sc->rx_ring_cnt; ++i)
4227                 igb_set_rxintr_mask(&sc->rx_rings[i], &intr_bit, intr_bitmax);
4228         sc->sts_intr_bit = 0;
4229         sc->sts_intr_mask = E1000_EICR_OTHER;
4230
4231         /* Initialize interrupt rate */
4232         sc->intr_rate = IGB_INTR_RATE;
4233 done:
4234         igb_set_ring_inuse(sc, FALSE);
4235         igb_set_intr_mask(sc);
4236         return 0;
4237 }
4238
4239 static void
4240 igb_free_intr(struct igb_softc *sc)
4241 {
4242         if (sc->intr_type != PCI_INTR_TYPE_MSIX) {
4243                 if (sc->intr_res != NULL) {
4244                         bus_release_resource(sc->dev, SYS_RES_IRQ, sc->intr_rid,
4245                             sc->intr_res);
4246                 }
4247                 if (sc->intr_type == PCI_INTR_TYPE_MSI)
4248                         pci_release_msi(sc->dev);
4249         } else {
4250                 igb_msix_free(sc, TRUE);
4251         }
4252 }
4253
4254 static void
4255 igb_teardown_intr(struct igb_softc *sc)
4256 {
4257         if (sc->intr_type != PCI_INTR_TYPE_MSIX)
4258                 bus_teardown_intr(sc->dev, sc->intr_res, sc->intr_tag);
4259         else
4260                 igb_msix_teardown(sc, sc->msix_cnt);
4261 }
4262
4263 static void
4264 igb_msix_try_alloc(struct igb_softc *sc)
4265 {
4266         int msix_enable, msix_cnt, msix_cnt2, alloc_cnt;
4267         int i, x, error;
4268         int offset, offset_def;
4269         struct igb_msix_data *msix;
4270         boolean_t aggregate, setup = FALSE;
4271
4272         /*
4273          * Don't enable MSI-X on 82575, see:
4274          * 82575 specification update errata #25
4275          */
4276         if (sc->hw.mac.type == e1000_82575)
4277                 return;
4278
4279         /* Don't enable MSI-X on VF */
4280         if (sc->vf_ifp)
4281                 return;
4282
4283         msix_enable = device_getenv_int(sc->dev, "msix.enable",
4284             igb_msix_enable);
4285         if (!msix_enable)
4286                 return;
4287
4288         msix_cnt = pci_msix_count(sc->dev);
4289 #ifdef IGB_MSIX_DEBUG
4290         msix_cnt = device_getenv_int(sc->dev, "msix.count", msix_cnt);
4291 #endif
4292         if (msix_cnt <= 1) {
4293                 /* One MSI-X model does not make sense */
4294                 return;
4295         }
4296
4297         i = 0;
4298         while ((1 << (i + 1)) <= msix_cnt)
4299                 ++i;
4300         msix_cnt2 = 1 << i;
4301
4302         if (bootverbose) {
4303                 device_printf(sc->dev, "MSI-X count %d/%d\n",
4304                     msix_cnt2, msix_cnt);
4305         }
4306
4307         KKASSERT(msix_cnt2 <= msix_cnt);
4308         if (msix_cnt == msix_cnt2) {
4309                 /* We need at least one MSI-X for link status */
4310                 msix_cnt2 >>= 1;
4311                 if (msix_cnt2 <= 1) {
4312                         /* One MSI-X for RX/TX does not make sense */
4313                         device_printf(sc->dev, "not enough MSI-X for TX/RX, "
4314                             "MSI-X count %d/%d\n", msix_cnt2, msix_cnt);
4315                         return;
4316                 }
4317                 KKASSERT(msix_cnt > msix_cnt2);
4318
4319                 if (bootverbose) {
4320                         device_printf(sc->dev, "MSI-X count fixup %d/%d\n",
4321                             msix_cnt2, msix_cnt);
4322                 }
4323         }
4324
4325         sc->rx_ring_msix = sc->rx_ring_cnt;
4326         if (sc->rx_ring_msix > msix_cnt2)
4327                 sc->rx_ring_msix = msix_cnt2;
4328
4329         sc->tx_ring_msix = sc->tx_ring_cnt;
4330         if (sc->tx_ring_msix > msix_cnt2)
4331                 sc->tx_ring_msix = msix_cnt2;
4332
4333         if (msix_cnt >= sc->tx_ring_msix + sc->rx_ring_msix + 1) {
4334                 /*
4335                  * Independent TX/RX MSI-X
4336                  */
4337                 aggregate = FALSE;
4338                 if (bootverbose)
4339                         device_printf(sc->dev, "independent TX/RX MSI-X\n");
4340                 alloc_cnt = sc->tx_ring_msix + sc->rx_ring_msix;
4341         } else {
4342                 /*
4343                  * Aggregate TX/RX MSI-X
4344                  */
4345                 aggregate = TRUE;
4346                 if (bootverbose)
4347                         device_printf(sc->dev, "aggregate TX/RX MSI-X\n");
4348                 alloc_cnt = msix_cnt2;
4349                 if (alloc_cnt > ncpus2)
4350                         alloc_cnt = ncpus2;
4351                 if (sc->rx_ring_msix > alloc_cnt)
4352                         sc->rx_ring_msix = alloc_cnt;
4353                 if (sc->tx_ring_msix > alloc_cnt)
4354                         sc->tx_ring_msix = alloc_cnt;
4355         }
4356         ++alloc_cnt;    /* For link status */
4357
4358         if (bootverbose) {
4359                 device_printf(sc->dev, "MSI-X alloc %d, "
4360                     "RX ring %d, TX ring %d\n", alloc_cnt,
4361                     sc->rx_ring_msix, sc->tx_ring_msix);
4362         }
4363
4364         sc->msix_mem_rid = PCIR_BAR(IGB_MSIX_BAR);
4365         sc->msix_mem_res = bus_alloc_resource_any(sc->dev, SYS_RES_MEMORY,
4366             &sc->msix_mem_rid, RF_ACTIVE);
4367         if (sc->msix_mem_res == NULL) {
4368                 sc->msix_mem_rid = PCIR_BAR(IGB_MSIX_BAR_ALT);
4369                 sc->msix_mem_res = bus_alloc_resource_any(sc->dev, SYS_RES_MEMORY,
4370                     &sc->msix_mem_rid, RF_ACTIVE);
4371                 if (sc->msix_mem_res == NULL) {
4372                         device_printf(sc->dev, "Unable to map MSI-X table\n");
4373                         return;
4374                 }
4375         }
4376
4377         sc->msix_cnt = alloc_cnt;
4378         sc->msix_data = kmalloc_cachealign(
4379             sizeof(struct igb_msix_data) * sc->msix_cnt,
4380             M_DEVBUF, M_WAITOK | M_ZERO);
4381         for (x = 0; x < sc->msix_cnt; ++x) {
4382                 msix = &sc->msix_data[x];
4383
4384                 lwkt_serialize_init(&msix->msix_serialize0);
4385                 msix->msix_sc = sc;
4386                 msix->msix_rid = -1;
4387                 msix->msix_vector = x;
4388                 msix->msix_mask = 1 << msix->msix_vector;
4389                 msix->msix_rate = IGB_INTR_RATE;
4390         }
4391
4392         x = 0;
4393         if (!aggregate) {
4394                 /*
4395                  * RX rings
4396                  */
4397                 if (sc->rx_ring_msix == ncpus2) {
4398                         offset = 0;
4399                 } else {
4400                         offset_def = (sc->rx_ring_msix *
4401                             device_get_unit(sc->dev)) % ncpus2;
4402
4403                         offset = device_getenv_int(sc->dev,
4404                             "msix.rxoff", offset_def);
4405                         if (offset >= ncpus2 ||
4406                             offset % sc->rx_ring_msix != 0) {
4407                                 device_printf(sc->dev,
4408                                     "invalid msix.rxoff %d, use %d\n",
4409                                     offset, offset_def);
4410                                 offset = offset_def;
4411                         }
4412                 }
4413                 igb_msix_rx_conf(sc, 0, &x, offset);
4414
4415                 /*
4416                  * TX rings
4417                  */
4418                 if (sc->tx_ring_msix == ncpus2) {
4419                         offset = 0;
4420                 } else {
4421                         offset_def = (sc->tx_ring_msix *
4422                             device_get_unit(sc->dev)) % ncpus2;
4423
4424                         offset = device_getenv_int(sc->dev,
4425                             "msix.txoff", offset_def);
4426                         if (offset >= ncpus2 ||
4427                             offset % sc->tx_ring_msix != 0) {
4428                                 device_printf(sc->dev,
4429                                     "invalid msix.txoff %d, use %d\n",
4430                                     offset, offset_def);
4431                                 offset = offset_def;
4432                         }
4433                 }
4434                 igb_msix_tx_conf(sc, 0, &x, offset);
4435         } else {
4436                 int ring_agg, ring_max;
4437
4438                 ring_agg = sc->rx_ring_msix;
4439                 if (ring_agg > sc->tx_ring_msix)
4440                         ring_agg = sc->tx_ring_msix;
4441
4442                 ring_max = sc->rx_ring_msix;
4443                 if (ring_max < sc->tx_ring_msix)
4444                         ring_max = sc->tx_ring_msix;
4445
4446                 if (ring_max == ncpus2) {
4447                         offset = 0;
4448                 } else {
4449                         offset_def = (ring_max * device_get_unit(sc->dev)) %
4450                             ncpus2;
4451
4452                         offset = device_getenv_int(sc->dev, "msix.off",
4453                             offset_def);
4454                         if (offset >= ncpus2 || offset % ring_max != 0) {
4455                                 device_printf(sc->dev,
4456                                     "invalid msix.off %d, use %d\n",
4457                                     offset, offset_def);
4458                                 offset = offset_def;
4459                         }
4460                 }
4461
4462                 for (i = 0; i < ring_agg; ++i) {
4463                         struct igb_tx_ring *txr = &sc->tx_rings[i];
4464                         struct igb_rx_ring *rxr = &sc->rx_rings[i];
4465
4466                         KKASSERT(x < sc->msix_cnt);
4467                         msix = &sc->msix_data[x++];
4468
4469                         txr->tx_intr_bit = msix->msix_vector;
4470                         txr->tx_intr_mask = msix->msix_mask;
4471                         rxr->rx_intr_bit = msix->msix_vector;
4472                         rxr->rx_intr_mask = msix->msix_mask;
4473
4474                         msix->msix_serialize = &msix->msix_serialize0;
4475                         msix->msix_func = igb_msix_rxtx;
4476                         msix->msix_arg = msix;
4477                         msix->msix_rx = rxr;
4478                         msix->msix_tx = txr;
4479
4480                         msix->msix_cpuid = i + offset;
4481                         KKASSERT(msix->msix_cpuid < ncpus2);
4482                         txr->tx_intr_cpuid = msix->msix_cpuid;
4483
4484                         ksnprintf(msix->msix_desc, sizeof(msix->msix_desc),
4485                             "%s rxtx%d", device_get_nameunit(sc->dev), i);
4486                         msix->msix_rate = IGB_MSIX_RX_RATE;
4487                         ksnprintf(msix->msix_rate_desc,
4488                             sizeof(msix->msix_rate_desc),
4489                             "RXTX%d interrupt rate", i);
4490                 }
4491
4492                 if (ring_agg != ring_max) {
4493                         if (ring_max == sc->tx_ring_msix)
4494                                 igb_msix_tx_conf(sc, i, &x, offset);
4495                         else
4496                                 igb_msix_rx_conf(sc, i, &x, offset);
4497                 }
4498         }
4499
4500         /*
4501          * Link status
4502          */
4503         KKASSERT(x < sc->msix_cnt);
4504         msix = &sc->msix_data[x++];
4505         sc->sts_intr_bit = msix->msix_vector;
4506         sc->sts_intr_mask = msix->msix_mask;
4507
4508         msix->msix_serialize = &sc->main_serialize;
4509         msix->msix_func = igb_msix_status;
4510         msix->msix_arg = sc;
4511         msix->msix_cpuid = 0;
4512         ksnprintf(msix->msix_desc, sizeof(msix->msix_desc), "%s sts",
4513             device_get_nameunit(sc->dev));
4514         ksnprintf(msix->msix_rate_desc, sizeof(msix->msix_rate_desc),
4515             "status interrupt rate");
4516
4517         KKASSERT(x == sc->msix_cnt);
4518
4519         error = pci_setup_msix(sc->dev);
4520         if (error) {
4521                 device_printf(sc->dev, "Setup MSI-X failed\n");
4522                 goto back;
4523         }
4524         setup = TRUE;
4525
4526         for (i = 0; i < sc->msix_cnt; ++i) {
4527                 msix = &sc->msix_data[i];
4528
4529                 error = pci_alloc_msix_vector(sc->dev, msix->msix_vector,
4530                     &msix->msix_rid, msix->msix_cpuid);
4531                 if (error) {
4532                         device_printf(sc->dev,
4533                             "Unable to allocate MSI-X %d on cpu%d\n",
4534                             msix->msix_vector, msix->msix_cpuid);
4535                         goto back;
4536                 }
4537
4538                 msix->msix_res = bus_alloc_resource_any(sc->dev, SYS_RES_IRQ,
4539                     &msix->msix_rid, RF_ACTIVE);
4540                 if (msix->msix_res == NULL) {
4541                         device_printf(sc->dev,
4542                             "Unable to allocate MSI-X %d resource\n",
4543                             msix->msix_vector);
4544                         error = ENOMEM;
4545                         goto back;
4546                 }
4547         }
4548
4549         pci_enable_msix(sc->dev);
4550         sc->intr_type = PCI_INTR_TYPE_MSIX;
4551 back:
4552         if (error)
4553                 igb_msix_free(sc, setup);
4554 }
4555
4556 static void
4557 igb_msix_free(struct igb_softc *sc, boolean_t setup)
4558 {
4559         int i;
4560
4561         KKASSERT(sc->msix_cnt > 1);
4562
4563         for (i = 0; i < sc->msix_cnt; ++i) {
4564                 struct igb_msix_data *msix = &sc->msix_data[i];
4565
4566                 if (msix->msix_res != NULL) {
4567                         bus_release_resource(sc->dev, SYS_RES_IRQ,
4568                             msix->msix_rid, msix->msix_res);
4569                 }
4570                 if (msix->msix_rid >= 0)
4571                         pci_release_msix_vector(sc->dev, msix->msix_rid);
4572         }
4573         if (setup)
4574                 pci_teardown_msix(sc->dev);
4575
4576         sc->msix_cnt = 0;
4577         kfree(sc->msix_data, M_DEVBUF);
4578         sc->msix_data = NULL;
4579 }
4580
4581 static int
4582 igb_msix_setup(struct igb_softc *sc)
4583 {
4584         int i;
4585
4586         for (i = 0; i < sc->msix_cnt; ++i) {
4587                 struct igb_msix_data *msix = &sc->msix_data[i];
4588                 int error;
4589
4590                 error = bus_setup_intr_descr(sc->dev, msix->msix_res,
4591                     INTR_MPSAFE, msix->msix_func, msix->msix_arg,
4592                     &msix->msix_handle, msix->msix_serialize, msix->msix_desc);
4593                 if (error) {
4594                         device_printf(sc->dev, "could not set up %s "
4595                             "interrupt handler.\n", msix->msix_desc);
4596                         igb_msix_teardown(sc, i);
4597                         return error;
4598                 }
4599         }
4600         return 0;
4601 }
4602
4603 static void
4604 igb_msix_teardown(struct igb_softc *sc, int msix_cnt)
4605 {
4606         int i;
4607
4608         for (i = 0; i < msix_cnt; ++i) {
4609                 struct igb_msix_data *msix = &sc->msix_data[i];
4610
4611                 bus_teardown_intr(sc->dev, msix->msix_res, msix->msix_handle);
4612         }
4613 }
4614
4615 static void
4616 igb_msix_rx(void *arg)
4617 {
4618         struct igb_rx_ring *rxr = arg;
4619
4620         ASSERT_SERIALIZED(&rxr->rx_serialize);
4621         igb_rxeof(rxr, -1);
4622
4623         E1000_WRITE_REG(&rxr->sc->hw, E1000_EIMS, rxr->rx_intr_mask);
4624 }
4625
4626 static void
4627 igb_msix_tx(void *arg)
4628 {
4629         struct igb_tx_ring *txr = arg;
4630
4631         ASSERT_SERIALIZED(&txr->tx_serialize);
4632
4633         igb_txeof(txr);
4634         if (!ifsq_is_empty(txr->ifsq))
4635                 ifsq_devstart(txr->ifsq);
4636
4637         E1000_WRITE_REG(&txr->sc->hw, E1000_EIMS, txr->tx_intr_mask);
4638 }
4639
4640 static void
4641 igb_msix_status(void *arg)
4642 {
4643         struct igb_softc *sc = arg;
4644         uint32_t icr;
4645
4646         ASSERT_SERIALIZED(&sc->main_serialize);
4647
4648         icr = E1000_READ_REG(&sc->hw, E1000_ICR);
4649         if (icr & E1000_ICR_LSC) {
4650                 sc->hw.mac.get_link_status = 1;
4651                 igb_update_link_status(sc);
4652         }
4653
4654         E1000_WRITE_REG(&sc->hw, E1000_EIMS, sc->sts_intr_mask);
4655 }
4656
4657 static void
4658 igb_set_ring_inuse(struct igb_softc *sc, boolean_t polling)
4659 {
4660         sc->rx_ring_inuse = igb_get_rxring_inuse(sc, polling);
4661         sc->tx_ring_inuse = igb_get_txring_inuse(sc, polling);
4662         if (bootverbose) {
4663                 if_printf(&sc->arpcom.ac_if, "RX rings %d/%d, TX rings %d/%d\n",
4664                     sc->rx_ring_inuse, sc->rx_ring_cnt,
4665                     sc->tx_ring_inuse, sc->tx_ring_cnt);
4666         }
4667 }
4668
4669 static int
4670 igb_get_rxring_inuse(const struct igb_softc *sc, boolean_t polling)
4671 {
4672         if (!IGB_ENABLE_HWRSS(sc))
4673                 return 1;
4674
4675         if (polling)
4676                 return sc->rx_ring_cnt;
4677         else if (sc->intr_type != PCI_INTR_TYPE_MSIX)
4678                 return IGB_MIN_RING_RSS;
4679         else
4680                 return sc->rx_ring_msix;
4681 }
4682
4683 static int
4684 igb_get_txring_inuse(const struct igb_softc *sc, boolean_t polling)
4685 {
4686         if (!IGB_ENABLE_HWTSS(sc))
4687                 return 1;
4688
4689         if (polling)
4690                 return sc->tx_ring_cnt;
4691         else if (sc->intr_type != PCI_INTR_TYPE_MSIX)
4692                 return IGB_MIN_RING;
4693         else
4694                 return sc->tx_ring_msix;
4695 }
4696
4697 static int
4698 igb_tso_pullup(struct igb_tx_ring *txr, struct mbuf **mp)
4699 {
4700         int hoff, iphlen, thoff;
4701         struct mbuf *m;
4702
4703         m = *mp;
4704         KASSERT(M_WRITABLE(m), ("TSO mbuf not writable"));
4705
4706         iphlen = m->m_pkthdr.csum_iphlen;
4707         thoff = m->m_pkthdr.csum_thlen;
4708         hoff = m->m_pkthdr.csum_lhlen;
4709
4710         KASSERT(iphlen > 0, ("invalid ip hlen"));
4711         KASSERT(thoff > 0, ("invalid tcp hlen"));
4712         KASSERT(hoff > 0, ("invalid ether hlen"));
4713
4714         if (__predict_false(m->m_len < hoff + iphlen + thoff)) {
4715                 m = m_pullup(m, hoff + iphlen + thoff);
4716                 if (m == NULL) {
4717                         *mp = NULL;
4718                         return ENOBUFS;
4719                 }
4720                 *mp = m;
4721         }
4722         if (txr->tx_flags & IGB_TXFLAG_TSO_IPLEN0) {
4723                 struct ip *ip;
4724
4725                 ip = mtodoff(m, struct ip *, hoff);
4726                 ip->ip_len = 0;
4727         }
4728
4729         return 0;
4730 }
4731
4732 static void
4733 igb_tso_ctx(struct igb_tx_ring *txr, struct mbuf *m, uint32_t *hlen)
4734 {
4735         struct e1000_adv_tx_context_desc *TXD;
4736         uint32_t vlan_macip_lens, type_tucmd_mlhl, mss_l4len_idx;
4737         int hoff, ctxd, iphlen, thoff;
4738
4739         iphlen = m->m_pkthdr.csum_iphlen;
4740         thoff = m->m_pkthdr.csum_thlen;
4741         hoff = m->m_pkthdr.csum_lhlen;
4742
4743         vlan_macip_lens = type_tucmd_mlhl = mss_l4len_idx = 0;
4744
4745         ctxd = txr->next_avail_desc;
4746         TXD = (struct e1000_adv_tx_context_desc *)&txr->tx_base[ctxd];
4747
4748         if (m->m_flags & M_VLANTAG) {
4749                 uint16_t vlantag;
4750
4751                 vlantag = htole16(m->m_pkthdr.ether_vlantag);
4752                 vlan_macip_lens |= (vlantag << E1000_ADVTXD_VLAN_SHIFT);
4753         }
4754
4755         vlan_macip_lens |= (hoff << E1000_ADVTXD_MACLEN_SHIFT);
4756         vlan_macip_lens |= iphlen;
4757
4758         type_tucmd_mlhl |= E1000_ADVTXD_DCMD_DEXT | E1000_ADVTXD_DTYP_CTXT;
4759         type_tucmd_mlhl |= E1000_ADVTXD_TUCMD_L4T_TCP;
4760         type_tucmd_mlhl |= E1000_ADVTXD_TUCMD_IPV4;
4761
4762         mss_l4len_idx |= (m->m_pkthdr.tso_segsz << E1000_ADVTXD_MSS_SHIFT);
4763         mss_l4len_idx |= (thoff << E1000_ADVTXD_L4LEN_SHIFT);
4764
4765         /*
4766          * 82575 needs the TX context index added; the queue
4767          * index is used as TX context index here.
4768          */
4769         if (txr->sc->hw.mac.type == e1000_82575)
4770                 mss_l4len_idx |= txr->me << 4;
4771
4772         TXD->vlan_macip_lens = htole32(vlan_macip_lens);
4773         TXD->type_tucmd_mlhl = htole32(type_tucmd_mlhl);
4774         TXD->seqnum_seed = htole32(0);
4775         TXD->mss_l4len_idx = htole32(mss_l4len_idx);
4776
4777         /* We've consumed the first desc, adjust counters */
4778         if (++ctxd == txr->num_tx_desc)
4779                 ctxd = 0;
4780         txr->next_avail_desc = ctxd;
4781         --txr->tx_avail;
4782
4783         *hlen = hoff + iphlen + thoff;
4784 }
4785
4786 static void
4787 igb_setup_serializer(struct igb_softc *sc)
4788 {
4789         const struct igb_msix_data *msix;
4790         int i, j;
4791
4792         /*
4793          * Allocate serializer array
4794          */
4795
4796         /* Main + TX + RX */
4797         sc->serialize_cnt = 1 + sc->tx_ring_cnt + sc->rx_ring_cnt;
4798
4799         /* Aggregate TX/RX MSI-X */
4800         for (i = 0; i < sc->msix_cnt; ++i) {
4801                 msix = &sc->msix_data[i];
4802                 if (msix->msix_serialize == &msix->msix_serialize0)
4803                         sc->serialize_cnt++;
4804         }
4805
4806         sc->serializes =
4807             kmalloc(sc->serialize_cnt * sizeof(struct lwkt_serialize *),
4808                 M_DEVBUF, M_WAITOK | M_ZERO);
4809
4810         /*
4811          * Setup serializers
4812          *
4813          * NOTE: Order is critical
4814          */
4815
4816         i = 0;
4817
4818         KKASSERT(i < sc->serialize_cnt);
4819         sc->serializes[i++] = &sc->main_serialize;
4820
4821         for (j = 0; j < sc->msix_cnt; ++j) {
4822                 msix = &sc->msix_data[j];
4823                 if (msix->msix_serialize == &msix->msix_serialize0) {
4824                         KKASSERT(i < sc->serialize_cnt);
4825                         sc->serializes[i++] = msix->msix_serialize;
4826                 }
4827         }
4828
4829         for (j = 0; j < sc->tx_ring_cnt; ++j) {
4830                 KKASSERT(i < sc->serialize_cnt);
4831                 sc->serializes[i++] = &sc->tx_rings[j].tx_serialize;
4832         }
4833
4834         for (j = 0; j < sc->rx_ring_cnt; ++j) {
4835                 KKASSERT(i < sc->serialize_cnt);
4836                 sc->serializes[i++] = &sc->rx_rings[j].rx_serialize;
4837         }
4838
4839         KKASSERT(i == sc->serialize_cnt);
4840 }
4841
4842 static void
4843 igb_msix_rx_conf(struct igb_softc *sc, int i, int *x0, int offset)
4844 {
4845         int x = *x0;
4846
4847         for (; i < sc->rx_ring_msix; ++i) {
4848                 struct igb_rx_ring *rxr = &sc->rx_rings[i];
4849                 struct igb_msix_data *msix;
4850
4851                 KKASSERT(x < sc->msix_cnt);
4852                 msix = &sc->msix_data[x++];
4853
4854                 rxr->rx_intr_bit = msix->msix_vector;
4855                 rxr->rx_intr_mask = msix->msix_mask;
4856
4857                 msix->msix_serialize = &rxr->rx_serialize;
4858                 msix->msix_func = igb_msix_rx;
4859                 msix->msix_arg = rxr;
4860
4861                 msix->msix_cpuid = i + offset;
4862                 KKASSERT(msix->msix_cpuid < ncpus2);
4863
4864                 ksnprintf(msix->msix_desc, sizeof(msix->msix_desc), "%s rx%d",
4865                     device_get_nameunit(sc->dev), i);
4866
4867                 msix->msix_rate = IGB_MSIX_RX_RATE;
4868                 ksnprintf(msix->msix_rate_desc, sizeof(msix->msix_rate_desc),
4869                     "RX%d interrupt rate", i);
4870         }
4871         *x0 = x;
4872 }
4873
4874 static void
4875 igb_msix_tx_conf(struct igb_softc *sc, int i, int *x0, int offset)
4876 {
4877         int x = *x0;
4878
4879         for (; i < sc->tx_ring_msix; ++i) {
4880                 struct igb_tx_ring *txr = &sc->tx_rings[i];
4881                 struct igb_msix_data *msix;
4882
4883                 KKASSERT(x < sc->msix_cnt);
4884                 msix = &sc->msix_data[x++];
4885
4886                 txr->tx_intr_bit = msix->msix_vector;
4887                 txr->tx_intr_mask = msix->msix_mask;
4888
4889                 msix->msix_serialize = &txr->tx_serialize;
4890                 msix->msix_func = igb_msix_tx;
4891                 msix->msix_arg = txr;
4892
4893                 msix->msix_cpuid = i + offset;
4894                 KKASSERT(msix->msix_cpuid < ncpus2);
4895                 txr->tx_intr_cpuid = msix->msix_cpuid;
4896
4897                 ksnprintf(msix->msix_desc, sizeof(msix->msix_desc), "%s tx%d",
4898                     device_get_nameunit(sc->dev), i);
4899
4900                 msix->msix_rate = IGB_MSIX_TX_RATE;
4901                 ksnprintf(msix->msix_rate_desc, sizeof(msix->msix_rate_desc),
4902                     "TX%d interrupt rate", i);
4903         }
4904         *x0 = x;
4905 }
4906
4907 static void
4908 igb_msix_rxtx(void *arg)
4909 {
4910         struct igb_msix_data *msix = arg;
4911         struct igb_rx_ring *rxr = msix->msix_rx;
4912         struct igb_tx_ring *txr = msix->msix_tx;
4913
4914         ASSERT_SERIALIZED(&msix->msix_serialize0);
4915
4916         lwkt_serialize_enter(&rxr->rx_serialize);
4917         igb_rxeof(rxr, -1);
4918         lwkt_serialize_exit(&rxr->rx_serialize);
4919
4920         lwkt_serialize_enter(&txr->tx_serialize);
4921         igb_txeof(txr);
4922         if (!ifsq_is_empty(txr->ifsq))
4923                 ifsq_devstart(txr->ifsq);
4924         lwkt_serialize_exit(&txr->tx_serialize);
4925
4926         E1000_WRITE_REG(&msix->msix_sc->hw, E1000_EIMS, msix->msix_mask);
4927 }
4928
4929 static void
4930 igb_set_timer_cpuid(struct igb_softc *sc, boolean_t polling)
4931 {
4932         if (polling || sc->intr_type == PCI_INTR_TYPE_MSIX)
4933                 sc->timer_cpuid = 0; /* XXX fixed */
4934         else
4935                 sc->timer_cpuid = rman_get_cpuid(sc->intr_res);
4936 }