/*
 * Core code for QEMU igb emulation
 *
 * Datasheet:
 * https://www.intel.com/content/dam/www/public/us/en/documents/datasheets/82576eg-gbe-datasheet.pdf
 *
 * Copyright (c) 2020-2023 Red Hat, Inc.
 * Copyright (c) 2015 Ravello Systems LTD (http://ravellosystems.com)
 * Developed by Daynix Computing LTD (http://www.daynix.com)
 *
 * Authors:
 * Akihiko Odaki <akihiko.odaki@daynix.com>
 * Gal Hammmer <gal.hammer@sap.com>
 * Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
 * Dmitry Fleytman <dmitry@daynix.com>
 * Leonid Bloch <leonid@daynix.com>
 * Yan Vugenfirer <yan@daynix.com>
 *
 * Based on work done by:
 * Nir Peleg, Tutis Systems Ltd. for Qumranet Inc.
 * Copyright (c) 2008 Qumranet
 * Based on work done by:
 * Copyright (c) 2007 Dan Aloni
 * Copyright (c) 2004 Antony T Curtis
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, see <http://www.gnu.org/licenses/>.
 */

#include "qemu/osdep.h"
#include "qemu/log.h"
#include "net/net.h"
#include "net/tap.h"
#include "hw/net/mii.h"
#include "hw/pci/msi.h"
#include "hw/pci/msix.h"
#include "sysemu/runstate.h"

#include "net_tx_pkt.h"
#include "net_rx_pkt.h"

#include "igb_common.h"
#include "e1000x_common.h"
#include "igb_core.h"

#include "trace.h"

#define E1000E_MAX_TX_FRAGS (64)

union e1000_rx_desc_union {
    struct e1000_rx_desc legacy;
    union e1000_adv_rx_desc adv;
};

typedef struct IGBTxPktVmdqCallbackContext {
    IGBCore *core;
    NetClientState *nc;
} IGBTxPktVmdqCallbackContext;

typedef struct L2Header {
    struct eth_header eth;
    struct vlan_header vlan[2];
} L2Header;

typedef struct PTP2 {
    uint8_t message_id_transport_specific;
    uint8_t version_ptp;
    uint16_t message_length;
    uint8_t subdomain_number;
    uint8_t reserved0;
    uint16_t flags;
    uint64_t correction;
    uint8_t reserved1[5];
    uint8_t source_communication_technology;
    uint32_t source_uuid_lo;
    uint16_t source_uuid_hi;
    uint16_t source_port_id;
    uint16_t sequence_id;
    uint8_t control;
    uint8_t log_message_period;
} PTP2;

static ssize_t
igb_receive_internal(IGBCore *core, const struct iovec *iov, int iovcnt,
                     bool has_vnet, bool *external_tx);

static void igb_raise_interrupts(IGBCore *core, size_t index, uint32_t causes);
static void igb_reset(IGBCore *core, bool sw);

static inline void
igb_raise_legacy_irq(IGBCore *core)
{
    trace_e1000e_irq_legacy_notify(true);
    e1000x_inc_reg_if_not_full(core->mac, IAC);
    pci_set_irq(core->owner, 1);
}

static inline void
igb_lower_legacy_irq(IGBCore *core)
{
    trace_e1000e_irq_legacy_notify(false);
    pci_set_irq(core->owner, 0);
}

static void igb_msix_notify(IGBCore *core, unsigned int cause)
{
    PCIDevice *dev = core->owner;
    uint16_t vfn;
    uint32_t effective_eiac;
    unsigned int vector;

    vfn = 8 - (cause + 2) / IGBVF_MSIX_VEC_NUM;
    if (vfn < pcie_sriov_num_vfs(core->owner)) {
        dev = pcie_sriov_get_vf_at_index(core->owner, vfn);
        assert(dev);
        vector = (cause + 2) % IGBVF_MSIX_VEC_NUM;
    } else if (cause >= IGB_MSIX_VEC_NUM) {
        qemu_log_mask(LOG_GUEST_ERROR,
                      "igb: Tried to use vector unavailable for PF");
        return;
    } else {
        vector = cause;
    }

    msix_notify(dev, vector);

    trace_e1000e_irq_icr_clear_eiac(core->mac[EICR], core->mac[EIAC]);
    effective_eiac = core->mac[EIAC] & BIT(cause);
    core->mac[EICR] &= ~effective_eiac;
}

static inline void
igb_intrmgr_rearm_timer(IGBIntrDelayTimer *timer)
{
    int64_t delay_ns = (int64_t) timer->core->mac[timer->delay_reg] *
                                 timer->delay_resolution_ns;

    trace_e1000e_irq_rearm_timer(timer->delay_reg << 2, delay_ns);

    timer_mod(timer->timer, qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + delay_ns);

    timer->running = true;
}

static void
igb_intmgr_timer_resume(IGBIntrDelayTimer *timer)
{
    if (timer->running) {
        igb_intrmgr_rearm_timer(timer);
    }
}

static void
igb_intmgr_timer_pause(IGBIntrDelayTimer *timer)
{
    if (timer->running) {
        timer_del(timer->timer);
    }
}

static void
igb_intrmgr_on_msix_throttling_timer(void *opaque)
{
    IGBIntrDelayTimer *timer = opaque;
    int idx = timer - &timer->core->eitr[0];

    timer->running = false;

    trace_e1000e_irq_msix_notify_postponed_vec(idx);
    igb_msix_notify(timer->core, idx);
}

static void
igb_intrmgr_initialize_all_timers(IGBCore *core, bool create)
{
    int i;

    for (i = 0; i < IGB_INTR_NUM; i++) {
        core->eitr[i].core = core;
        core->eitr[i].delay_reg = EITR0 + i;
        core->eitr[i].delay_resolution_ns = E1000_INTR_DELAY_NS_RES;
    }

    if (!create) {
        return;
    }

    for (i = 0; i < IGB_INTR_NUM; i++) {
        core->eitr[i].timer = timer_new_ns(QEMU_CLOCK_VIRTUAL,
                                           igb_intrmgr_on_msix_throttling_timer,
                                           &core->eitr[i]);
    }
}

static void
igb_intrmgr_resume(IGBCore *core)
{
    int i;

    for (i = 0; i < IGB_INTR_NUM; i++) {
        igb_intmgr_timer_resume(&core->eitr[i]);
    }
}

static void
igb_intrmgr_pause(IGBCore *core)
{
    int i;

    for (i = 0; i < IGB_INTR_NUM; i++) {
        igb_intmgr_timer_pause(&core->eitr[i]);
    }
}

static void
igb_intrmgr_reset(IGBCore *core)
{
    int i;

    for (i = 0; i < IGB_INTR_NUM; i++) {
        if (core->eitr[i].running) {
            timer_del(core->eitr[i].timer);
            igb_intrmgr_on_msix_throttling_timer(&core->eitr[i]);
        }
    }
}

static void
igb_intrmgr_pci_unint(IGBCore *core)
{
    int i;

    for (i = 0; i < IGB_INTR_NUM; i++) {
        timer_free(core->eitr[i].timer);
    }
}

static void
igb_intrmgr_pci_realize(IGBCore *core)
{
    igb_intrmgr_initialize_all_timers(core, true);
}

static inline bool
igb_rx_csum_enabled(IGBCore *core)
{
    return (core->mac[RXCSUM] & E1000_RXCSUM_PCSD) ? false : true;
}

static inline bool
igb_rx_use_legacy_descriptor(IGBCore *core)
{
    /*
     * TODO: If SRRCTL[n],DESCTYPE = 000b, the 82576 uses the legacy Rx
     * descriptor.
     */
    return false;
}

static inline bool
igb_rss_enabled(IGBCore *core)
{
    return (core->mac[MRQC] & 3) == E1000_MRQC_ENABLE_RSS_MQ &&
           !igb_rx_csum_enabled(core) &&
           !igb_rx_use_legacy_descriptor(core);
}

typedef struct E1000E_RSSInfo_st {
    bool enabled;
    uint32_t hash;
    uint32_t queue;
    uint32_t type;
} E1000E_RSSInfo;

static uint32_t
igb_rss_get_hash_type(IGBCore *core, struct NetRxPkt *pkt)
{
    bool hasip4, hasip6;
    EthL4HdrProto l4hdr_proto;

    assert(igb_rss_enabled(core));

    net_rx_pkt_get_protocols(pkt, &hasip4, &hasip6, &l4hdr_proto);

    if (hasip4) {
        trace_e1000e_rx_rss_ip4(l4hdr_proto, core->mac[MRQC],
                                E1000_MRQC_EN_TCPIPV4(core->mac[MRQC]),
                                E1000_MRQC_EN_IPV4(core->mac[MRQC]));

        if (l4hdr_proto == ETH_L4_HDR_PROTO_TCP &&
            E1000_MRQC_EN_TCPIPV4(core->mac[MRQC])) {
            return E1000_MRQ_RSS_TYPE_IPV4TCP;
        }

        if (l4hdr_proto == ETH_L4_HDR_PROTO_UDP &&
            (core->mac[MRQC] & E1000_MRQC_RSS_FIELD_IPV4_UDP)) {
            return E1000_MRQ_RSS_TYPE_IPV4UDP;
        }

        if (E1000_MRQC_EN_IPV4(core->mac[MRQC])) {
            return E1000_MRQ_RSS_TYPE_IPV4;
        }
    } else if (hasip6) {
        eth_ip6_hdr_info *ip6info = net_rx_pkt_get_ip6_info(pkt);

        bool ex_dis = core->mac[RFCTL] & E1000_RFCTL_IPV6_EX_DIS;
        bool new_ex_dis = core->mac[RFCTL] & E1000_RFCTL_NEW_IPV6_EXT_DIS;

        /*
         * Following two traces must not be combined because resulting
         * event will have 11 arguments totally and some trace backends
         * (at least "ust") have limitation of maximum 10 arguments per
         * event. Events with more arguments fail to compile for
         * backends like these.
         */
        trace_e1000e_rx_rss_ip6_rfctl(core->mac[RFCTL]);
        trace_e1000e_rx_rss_ip6(ex_dis, new_ex_dis, l4hdr_proto,
                                ip6info->has_ext_hdrs,
                                ip6info->rss_ex_dst_valid,
                                ip6info->rss_ex_src_valid,
                                core->mac[MRQC],
                                E1000_MRQC_EN_TCPIPV6EX(core->mac[MRQC]),
                                E1000_MRQC_EN_IPV6EX(core->mac[MRQC]),
                                E1000_MRQC_EN_IPV6(core->mac[MRQC]));

        if ((!ex_dis || !ip6info->has_ext_hdrs) &&
            (!new_ex_dis || !(ip6info->rss_ex_dst_valid ||
                              ip6info->rss_ex_src_valid))) {

            if (l4hdr_proto == ETH_L4_HDR_PROTO_TCP &&
                E1000_MRQC_EN_TCPIPV6EX(core->mac[MRQC])) {
                return E1000_MRQ_RSS_TYPE_IPV6TCPEX;
            }

            if (l4hdr_proto == ETH_L4_HDR_PROTO_UDP &&
                (core->mac[MRQC] & E1000_MRQC_RSS_FIELD_IPV6_UDP)) {
                return E1000_MRQ_RSS_TYPE_IPV6UDP;
            }

            if (E1000_MRQC_EN_IPV6EX(core->mac[MRQC])) {
                return E1000_MRQ_RSS_TYPE_IPV6EX;
            }

        }

        if (E1000_MRQC_EN_IPV6(core->mac[MRQC])) {
            return E1000_MRQ_RSS_TYPE_IPV6;
        }

    }

    return E1000_MRQ_RSS_TYPE_NONE;
}

static uint32_t
igb_rss_calc_hash(IGBCore *core, struct NetRxPkt *pkt, E1000E_RSSInfo *info)
{
    NetRxPktRssType type;

    assert(igb_rss_enabled(core));

    switch (info->type) {
    case E1000_MRQ_RSS_TYPE_IPV4:
        type = NetPktRssIpV4;
        break;
    case E1000_MRQ_RSS_TYPE_IPV4TCP:
        type = NetPktRssIpV4Tcp;
        break;
    case E1000_MRQ_RSS_TYPE_IPV6TCPEX:
        type = NetPktRssIpV6TcpEx;
        break;
    case E1000_MRQ_RSS_TYPE_IPV6:
        type = NetPktRssIpV6;
        break;
    case E1000_MRQ_RSS_TYPE_IPV6EX:
        type = NetPktRssIpV6Ex;
        break;
    case E1000_MRQ_RSS_TYPE_IPV4UDP:
        type = NetPktRssIpV4Udp;
        break;
    case E1000_MRQ_RSS_TYPE_IPV6UDP:
        type = NetPktRssIpV6Udp;
        break;
    default:
        assert(false);
        return 0;
    }

    return net_rx_pkt_calc_rss_hash(pkt, type, (uint8_t *) &core->mac[RSSRK]);
}

static void
igb_rss_parse_packet(IGBCore *core, struct NetRxPkt *pkt, bool tx,
                     E1000E_RSSInfo *info)
{
    trace_e1000e_rx_rss_started();

    if (tx || !igb_rss_enabled(core)) {
        info->enabled = false;
        info->hash = 0;
        info->queue = 0;
        info->type = 0;
        trace_e1000e_rx_rss_disabled();
        return;
    }

    info->enabled = true;

    info->type = igb_rss_get_hash_type(core, pkt);

    trace_e1000e_rx_rss_type(info->type);

    if (info->type == E1000_MRQ_RSS_TYPE_NONE) {
        info->hash = 0;
        info->queue = 0;
        return;
    }

    info->hash = igb_rss_calc_hash(core, pkt, info);
    info->queue = E1000_RSS_QUEUE(&core->mac[RETA], info->hash);
}

static void
igb_tx_insert_vlan(IGBCore *core, uint16_t qn, struct igb_tx *tx,
    uint16_t vlan, bool insert_vlan)
{
    if (core->mac[MRQC] & 1) {
        uint16_t pool = qn % IGB_NUM_VM_POOLS;

        if (core->mac[VMVIR0 + pool] & E1000_VMVIR_VLANA_DEFAULT) {
            /* always insert default VLAN */
            insert_vlan = true;
            vlan = core->mac[VMVIR0 + pool] & 0xffff;
        } else if (core->mac[VMVIR0 + pool] & E1000_VMVIR_VLANA_NEVER) {
            insert_vlan = false;
        }
    }

    if (insert_vlan) {
        net_tx_pkt_setup_vlan_header_ex(tx->tx_pkt, vlan,
            core->mac[VET] & 0xffff);
    }
}

static bool
igb_setup_tx_offloads(IGBCore *core, struct igb_tx *tx)
{
    uint32_t idx = (tx->first_olinfo_status >> 4) & 1;

    if (tx->first_cmd_type_len & E1000_ADVTXD_DCMD_TSE) {
        uint32_t mss = tx->ctx[idx].mss_l4len_idx >> E1000_ADVTXD_MSS_SHIFT;
        if (!net_tx_pkt_build_vheader(tx->tx_pkt, true, true, mss)) {
            return false;
        }

        net_tx_pkt_update_ip_checksums(tx->tx_pkt);
        e1000x_inc_reg_if_not_full(core->mac, TSCTC);
        return true;
    }

    if ((tx->first_olinfo_status & E1000_ADVTXD_POTS_TXSM) &&
        !((tx->ctx[idx].type_tucmd_mlhl & E1000_ADVTXD_TUCMD_L4T_SCTP) ?
          net_tx_pkt_update_sctp_checksum(tx->tx_pkt) :
          net_tx_pkt_build_vheader(tx->tx_pkt, false, true, 0))) {
        return false;
    }

    if (tx->first_olinfo_status & E1000_ADVTXD_POTS_IXSM) {
        net_tx_pkt_update_ip_hdr_checksum(tx->tx_pkt);
    }

    return true;
}

static void igb_tx_pkt_mac_callback(void *core,
                                    const struct iovec *iov,
                                    int iovcnt,
                                    const struct iovec *virt_iov,
                                    int virt_iovcnt)
{
    igb_receive_internal(core, virt_iov, virt_iovcnt, true, NULL);
}

static void igb_tx_pkt_vmdq_callback(void *opaque,
                                     const struct iovec *iov,
                                     int iovcnt,
                                     const struct iovec *virt_iov,
                                     int virt_iovcnt)
{
    IGBTxPktVmdqCallbackContext *context = opaque;
    bool external_tx;

    igb_receive_internal(context->core, virt_iov, virt_iovcnt, true,
                         &external_tx);

    if (external_tx) {
        if (context->core->has_vnet) {
            qemu_sendv_packet(context->nc, virt_iov, virt_iovcnt);
        } else {
            qemu_sendv_packet(context->nc, iov, iovcnt);
        }
    }
}

/* TX Packets Switching (7.10.3.6) */
static bool igb_tx_pkt_switch(IGBCore *core, struct igb_tx *tx,
                              NetClientState *nc)
{
    IGBTxPktVmdqCallbackContext context;

    /* TX switching is only used to serve VM to VM traffic. */
    if (!(core->mac[MRQC] & 1)) {
        goto send_out;
    }

    /* TX switching requires DTXSWC.Loopback_en bit enabled. */
    if (!(core->mac[DTXSWC] & E1000_DTXSWC_VMDQ_LOOPBACK_EN)) {
        goto send_out;
    }

    context.core = core;
    context.nc = nc;

    return net_tx_pkt_send_custom(tx->tx_pkt, false,
                                  igb_tx_pkt_vmdq_callback, &context);

send_out:
    return net_tx_pkt_send(tx->tx_pkt, nc);
}

static bool
igb_tx_pkt_send(IGBCore *core, struct igb_tx *tx, int queue_index)
{
    int target_queue = MIN(core->max_queue_num, queue_index);
    NetClientState *queue = qemu_get_subqueue(core->owner_nic, target_queue);

    if (!igb_setup_tx_offloads(core, tx)) {
        return false;
    }

    net_tx_pkt_dump(tx->tx_pkt);

    if ((core->phy[MII_BMCR] & MII_BMCR_LOOPBACK) ||
        ((core->mac[RCTL] & E1000_RCTL_LBM_MAC) == E1000_RCTL_LBM_MAC)) {
        return net_tx_pkt_send_custom(tx->tx_pkt, false,
                                      igb_tx_pkt_mac_callback, core);
    } else {
        return igb_tx_pkt_switch(core, tx, queue);
    }
}

static void
igb_on_tx_done_update_stats(IGBCore *core, struct NetTxPkt *tx_pkt, int qn)
{
    static const int PTCregs[6] = { PTC64, PTC127, PTC255, PTC511,
                                    PTC1023, PTC1522 };

    size_t tot_len = net_tx_pkt_get_total_len(tx_pkt) + 4;

    e1000x_increase_size_stats(core->mac, PTCregs, tot_len);
    e1000x_inc_reg_if_not_full(core->mac, TPT);
    e1000x_grow_8reg_if_not_full(core->mac, TOTL, tot_len);

    switch (net_tx_pkt_get_packet_type(tx_pkt)) {
    case ETH_PKT_BCAST:
        e1000x_inc_reg_if_not_full(core->mac, BPTC);
        break;
    case ETH_PKT_MCAST:
        e1000x_inc_reg_if_not_full(core->mac, MPTC);
        break;
    case ETH_PKT_UCAST:
        break;
    default:
        g_assert_not_reached();
    }

    e1000x_inc_reg_if_not_full(core->mac, GPTC);
    e1000x_grow_8reg_if_not_full(core->mac, GOTCL, tot_len);

    if (core->mac[MRQC] & 1) {
        uint16_t pool = qn % IGB_NUM_VM_POOLS;

        core->mac[PVFGOTC0 + (pool * 64)] += tot_len;
        core->mac[PVFGPTC0 + (pool * 64)]++;
    }
}

static void
igb_process_tx_desc(IGBCore *core,
                    PCIDevice *dev,
                    struct igb_tx *tx,
                    union e1000_adv_tx_desc *tx_desc,
                    int queue_index)
{
    struct e1000_adv_tx_context_desc *tx_ctx_desc;
    uint32_t cmd_type_len;
    uint32_t idx;
    uint64_t buffer_addr;
    uint16_t length;

    cmd_type_len = le32_to_cpu(tx_desc->read.cmd_type_len);

    if (cmd_type_len & E1000_ADVTXD_DCMD_DEXT) {
        if ((cmd_type_len & E1000_ADVTXD_DTYP_DATA) ==
            E1000_ADVTXD_DTYP_DATA) {
            /* advanced transmit data descriptor */
            if (tx->first) {
                tx->first_cmd_type_len = cmd_type_len;
                tx->first_olinfo_status = le32_to_cpu(tx_desc->read.olinfo_status);
                tx->first = false;
            }
        } else if ((cmd_type_len & E1000_ADVTXD_DTYP_CTXT) ==
                   E1000_ADVTXD_DTYP_CTXT) {
            /* advanced transmit context descriptor */
            tx_ctx_desc = (struct e1000_adv_tx_context_desc *)tx_desc;
            idx = (le32_to_cpu(tx_ctx_desc->mss_l4len_idx) >> 4) & 1;
            tx->ctx[idx].vlan_macip_lens = le32_to_cpu(tx_ctx_desc->vlan_macip_lens);
            tx->ctx[idx].seqnum_seed = le32_to_cpu(tx_ctx_desc->seqnum_seed);
            tx->ctx[idx].type_tucmd_mlhl = le32_to_cpu(tx_ctx_desc->type_tucmd_mlhl);
            tx->ctx[idx].mss_l4len_idx = le32_to_cpu(tx_ctx_desc->mss_l4len_idx);
            return;
        } else {
            /* unknown descriptor type */
            return;
        }
    } else {
        /* legacy descriptor */

        /* TODO: Implement a support for legacy descriptors (7.2.2.1). */
    }

    buffer_addr = le64_to_cpu(tx_desc->read.buffer_addr);
    length = cmd_type_len & 0xFFFF;

    if (!tx->skip_cp) {
        if (!net_tx_pkt_add_raw_fragment_pci(tx->tx_pkt, dev,
                                             buffer_addr, length)) {
            tx->skip_cp = true;
        }
    }

    if (cmd_type_len & E1000_TXD_CMD_EOP) {
        if (!tx->skip_cp && net_tx_pkt_parse(tx->tx_pkt)) {
            idx = (tx->first_olinfo_status >> 4) & 1;
            igb_tx_insert_vlan(core, queue_index, tx,
                tx->ctx[idx].vlan_macip_lens >> IGB_TX_FLAGS_VLAN_SHIFT,
                !!(tx->first_cmd_type_len & E1000_TXD_CMD_VLE));

            if ((tx->first_cmd_type_len & E1000_ADVTXD_MAC_TSTAMP) &&
                (core->mac[TSYNCTXCTL] & E1000_TSYNCTXCTL_ENABLED) &&
                !(core->mac[TSYNCTXCTL] & E1000_TSYNCTXCTL_VALID)) {
                core->mac[TSYNCTXCTL] |= E1000_TSYNCTXCTL_VALID;
                e1000x_timestamp(core->mac, core->timadj, TXSTMPL, TXSTMPH);
            }

            if (igb_tx_pkt_send(core, tx, queue_index)) {
                igb_on_tx_done_update_stats(core, tx->tx_pkt, queue_index);
            }
        }

        tx->first = true;
        tx->skip_cp = false;
        net_tx_pkt_reset(tx->tx_pkt, net_tx_pkt_unmap_frag_pci, dev);
    }
}

static uint32_t igb_tx_wb_eic(IGBCore *core, int queue_idx)
{
    uint32_t n, ent = 0;

    n = igb_ivar_entry_tx(queue_idx);
    ent = (core->mac[IVAR0 + n / 4] >> (8 * (n % 4))) & 0xff;

    return (ent & E1000_IVAR_VALID) ? BIT(ent & 0x1f) : 0;
}

static uint32_t igb_rx_wb_eic(IGBCore *core, int queue_idx)
{
    uint32_t n, ent = 0;

    n = igb_ivar_entry_rx(queue_idx);
    ent = (core->mac[IVAR0 + n / 4] >> (8 * (n % 4))) & 0xff;

    return (ent & E1000_IVAR_VALID) ? BIT(ent & 0x1f) : 0;
}

typedef struct E1000E_RingInfo_st {
    int dbah;
    int dbal;
    int dlen;
    int dh;
    int dt;
    int idx;
} E1000E_RingInfo;

static inline bool
igb_ring_empty(IGBCore *core, const E1000E_RingInfo *r)
{
    return core->mac[r->dh] == core->mac[r->dt] ||
                core->mac[r->dt] >= core->mac[r->dlen] / E1000_RING_DESC_LEN;
}

static inline uint64_t
igb_ring_base(IGBCore *core, const E1000E_RingInfo *r)
{
    uint64_t bah = core->mac[r->dbah];
    uint64_t bal = core->mac[r->dbal];

    return (bah << 32) + bal;
}

static inline uint64_t
igb_ring_head_descr(IGBCore *core, const E1000E_RingInfo *r)
{
    return igb_ring_base(core, r) + E1000_RING_DESC_LEN * core->mac[r->dh];
}

static inline void
igb_ring_advance(IGBCore *core, const E1000E_RingInfo *r, uint32_t count)
{
    core->mac[r->dh] += count;

    if (core->mac[r->dh] * E1000_RING_DESC_LEN >= core->mac[r->dlen]) {
        core->mac[r->dh] = 0;
    }
}

static inline uint32_t
igb_ring_free_descr_num(IGBCore *core, const E1000E_RingInfo *r)
{
    trace_e1000e_ring_free_space(r->idx, core->mac[r->dlen],
                                 core->mac[r->dh],  core->mac[r->dt]);

    if (core->mac[r->dh] <= core->mac[r->dt]) {
        return core->mac[r->dt] - core->mac[r->dh];
    }

    if (core->mac[r->dh] > core->mac[r->dt]) {
        return core->mac[r->dlen] / E1000_RING_DESC_LEN +
               core->mac[r->dt] - core->mac[r->dh];
    }

    g_assert_not_reached();
    return 0;
}

static inline bool
igb_ring_enabled(IGBCore *core, const E1000E_RingInfo *r)
{
    return core->mac[r->dlen] > 0;
}

typedef struct IGB_TxRing_st {
    const E1000E_RingInfo *i;
    struct igb_tx *tx;
} IGB_TxRing;

static inline int
igb_mq_queue_idx(int base_reg_idx, int reg_idx)
{
    return (reg_idx - base_reg_idx) / 16;
}

static inline void
igb_tx_ring_init(IGBCore *core, IGB_TxRing *txr, int idx)
{
    static const E1000E_RingInfo i[IGB_NUM_QUEUES] = {
        { TDBAH0, TDBAL0, TDLEN0, TDH0, TDT0, 0 },
        { TDBAH1, TDBAL1, TDLEN1, TDH1, TDT1, 1 },
        { TDBAH2, TDBAL2, TDLEN2, TDH2, TDT2, 2 },
        { TDBAH3, TDBAL3, TDLEN3, TDH3, TDT3, 3 },
        { TDBAH4, TDBAL4, TDLEN4, TDH4, TDT4, 4 },
        { TDBAH5, TDBAL5, TDLEN5, TDH5, TDT5, 5 },
        { TDBAH6, TDBAL6, TDLEN6, TDH6, TDT6, 6 },
        { TDBAH7, TDBAL7, TDLEN7, TDH7, TDT7, 7 },
        { TDBAH8, TDBAL8, TDLEN8, TDH8, TDT8, 8 },
        { TDBAH9, TDBAL9, TDLEN9, TDH9, TDT9, 9 },
        { TDBAH10, TDBAL10, TDLEN10, TDH10, TDT10, 10 },
        { TDBAH11, TDBAL11, TDLEN11, TDH11, TDT11, 11 },
        { TDBAH12, TDBAL12, TDLEN12, TDH12, TDT12, 12 },
        { TDBAH13, TDBAL13, TDLEN13, TDH13, TDT13, 13 },
        { TDBAH14, TDBAL14, TDLEN14, TDH14, TDT14, 14 },
        { TDBAH15, TDBAL15, TDLEN15, TDH15, TDT15, 15 }
    };

    assert(idx < ARRAY_SIZE(i));

    txr->i     = &i[idx];
    txr->tx    = &core->tx[idx];
}

typedef struct E1000E_RxRing_st {
    const E1000E_RingInfo *i;
} E1000E_RxRing;

static inline void
igb_rx_ring_init(IGBCore *core, E1000E_RxRing *rxr, int idx)
{
    static const E1000E_RingInfo i[IGB_NUM_QUEUES] = {
        { RDBAH0, RDBAL0, RDLEN0, RDH0, RDT0, 0 },
        { RDBAH1, RDBAL1, RDLEN1, RDH1, RDT1, 1 },
        { RDBAH2, RDBAL2, RDLEN2, RDH2, RDT2, 2 },
        { RDBAH3, RDBAL3, RDLEN3, RDH3, RDT3, 3 },
        { RDBAH4, RDBAL4, RDLEN4, RDH4, RDT4, 4 },
        { RDBAH5, RDBAL5, RDLEN5, RDH5, RDT5, 5 },
        { RDBAH6, RDBAL6, RDLEN6, RDH6, RDT6, 6 },
        { RDBAH7, RDBAL7, RDLEN7, RDH7, RDT7, 7 },
        { RDBAH8, RDBAL8, RDLEN8, RDH8, RDT8, 8 },
        { RDBAH9, RDBAL9, RDLEN9, RDH9, RDT9, 9 },
        { RDBAH10, RDBAL10, RDLEN10, RDH10, RDT10, 10 },
        { RDBAH11, RDBAL11, RDLEN11, RDH11, RDT11, 11 },
        { RDBAH12, RDBAL12, RDLEN12, RDH12, RDT12, 12 },
        { RDBAH13, RDBAL13, RDLEN13, RDH13, RDT13, 13 },
        { RDBAH14, RDBAL14, RDLEN14, RDH14, RDT14, 14 },
        { RDBAH15, RDBAL15, RDLEN15, RDH15, RDT15, 15 }
    };

    assert(idx < ARRAY_SIZE(i));

    rxr->i      = &i[idx];
}

static uint32_t
igb_txdesc_writeback(IGBCore *core, dma_addr_t base,
                     union e1000_adv_tx_desc *tx_desc,
                     const E1000E_RingInfo *txi)
{
    PCIDevice *d;
    uint32_t cmd_type_len = le32_to_cpu(tx_desc->read.cmd_type_len);
    uint64_t tdwba;

    tdwba = core->mac[E1000_TDWBAL(txi->idx) >> 2];
    tdwba |= (uint64_t)core->mac[E1000_TDWBAH(txi->idx) >> 2] << 32;

    if (!(cmd_type_len & E1000_TXD_CMD_RS)) {
        return 0;
    }

    d = pcie_sriov_get_vf_at_index(core->owner, txi->idx % 8);
    if (!d) {
        d = core->owner;
    }

    if (tdwba & 1) {
        uint32_t buffer = cpu_to_le32(core->mac[txi->dh]);
        pci_dma_write(d, tdwba & ~3, &buffer, sizeof(buffer));
    } else {
        uint32_t status = le32_to_cpu(tx_desc->wb.status) | E1000_TXD_STAT_DD;

        tx_desc->wb.status = cpu_to_le32(status);
        pci_dma_write(d, base + offsetof(union e1000_adv_tx_desc, wb),
            &tx_desc->wb, sizeof(tx_desc->wb));
    }

    return igb_tx_wb_eic(core, txi->idx);
}

static inline bool
igb_tx_enabled(IGBCore *core, const E1000E_RingInfo *txi)
{
    bool vmdq = core->mac[MRQC] & 1;
    uint16_t qn = txi->idx;
    uint16_t pool = qn % IGB_NUM_VM_POOLS;

    return (core->mac[TCTL] & E1000_TCTL_EN) &&
        (!vmdq || core->mac[VFTE] & BIT(pool)) &&
        (core->mac[TXDCTL0 + (qn * 16)] & E1000_TXDCTL_QUEUE_ENABLE);
}

static void
igb_start_xmit(IGBCore *core, const IGB_TxRing *txr)
{
    PCIDevice *d;
    dma_addr_t base;
    union e1000_adv_tx_desc desc;
    const E1000E_RingInfo *txi = txr->i;
    uint32_t eic = 0;

    if (!igb_tx_enabled(core, txi)) {
        trace_e1000e_tx_disabled();
        return;
    }

    d = pcie_sriov_get_vf_at_index(core->owner, txi->idx % 8);
    if (!d) {
        d = core->owner;
    }

    while (!igb_ring_empty(core, txi)) {
        base = igb_ring_head_descr(core, txi);

        pci_dma_read(d, base, &desc, sizeof(desc));

        trace_e1000e_tx_descr((void *)(intptr_t)desc.read.buffer_addr,
                              desc.read.cmd_type_len, desc.wb.status);

        igb_process_tx_desc(core, d, txr->tx, &desc, txi->idx);
        igb_ring_advance(core, txi, 1);
        eic |= igb_txdesc_writeback(core, base, &desc, txi);
    }

    if (eic) {
        igb_raise_interrupts(core, EICR, eic);
        igb_raise_interrupts(core, ICR, E1000_ICR_TXDW);
    }

    net_tx_pkt_reset(txr->tx->tx_pkt, net_tx_pkt_unmap_frag_pci, d);
}

static uint32_t
igb_rxbufsize(IGBCore *core, const E1000E_RingInfo *r)
{
    uint32_t srrctl = core->mac[E1000_SRRCTL(r->idx) >> 2];
    uint32_t bsizepkt = srrctl & E1000_SRRCTL_BSIZEPKT_MASK;
    if (bsizepkt) {
        return bsizepkt << E1000_SRRCTL_BSIZEPKT_SHIFT;
    }

    return e1000x_rxbufsize(core->mac[RCTL]);
}

static bool
igb_has_rxbufs(IGBCore *core, const E1000E_RingInfo *r, size_t total_size)
{
    uint32_t bufs = igb_ring_free_descr_num(core, r);
    uint32_t bufsize = igb_rxbufsize(core, r);

    trace_e1000e_rx_has_buffers(r->idx, bufs, total_size, bufsize);

    return total_size <= bufs / (core->rx_desc_len / E1000_MIN_RX_DESC_LEN) *
                         bufsize;
}

void
igb_start_recv(IGBCore *core)
{
    int i;

    trace_e1000e_rx_start_recv();

    for (i = 0; i <= core->max_queue_num; i++) {
        qemu_flush_queued_packets(qemu_get_subqueue(core->owner_nic, i));
    }
}

bool
igb_can_receive(IGBCore *core)
{
    int i;

    if (!e1000x_rx_ready(core->owner, core->mac)) {
        return false;
    }

    for (i = 0; i < IGB_NUM_QUEUES; i++) {
        E1000E_RxRing rxr;
        if (!(core->mac[RXDCTL0 + (i * 16)] & E1000_RXDCTL_QUEUE_ENABLE)) {
            continue;
        }

        igb_rx_ring_init(core, &rxr, i);
        if (igb_ring_enabled(core, rxr.i) && igb_has_rxbufs(core, rxr.i, 1)) {
            trace_e1000e_rx_can_recv();
            return true;
        }
    }

    trace_e1000e_rx_can_recv_rings_full();
    return false;
}

ssize_t
igb_receive(IGBCore *core, const uint8_t *buf, size_t size)
{
    const struct iovec iov = {
        .iov_base = (uint8_t *)buf,
        .iov_len = size
    };

    return igb_receive_iov(core, &iov, 1);
}

static inline bool
igb_rx_l3_cso_enabled(IGBCore *core)
{
    return !!(core->mac[RXCSUM] & E1000_RXCSUM_IPOFLD);
}

static inline bool
igb_rx_l4_cso_enabled(IGBCore *core)

{
    return !!(core->mac[RXCSUM] & E1000_RXCSUM_TUOFLD);
}

static bool igb_rx_is_oversized(IGBCore *core, const struct eth_header *ehdr,
                                size_t size, size_t vlan_num,
                                bool lpe, uint16_t rlpml)
{
    size_t vlan_header_size = sizeof(struct vlan_header) * vlan_num;
    size_t header_size = sizeof(struct eth_header) + vlan_header_size;
    return lpe ? size + ETH_FCS_LEN > rlpml : size > header_size + ETH_MTU;
}

static uint16_t igb_receive_assign(IGBCore *core, const struct iovec *iov,
                                   size_t iovcnt, size_t iov_ofs,
                                   const L2Header *l2_header, size_t size,
                                   E1000E_RSSInfo *rss_info,
                                   uint16_t *etqf, bool *ts, bool *external_tx)
{
    static const int ta_shift[] = { 4, 3, 2, 0 };
    const struct eth_header *ehdr = &l2_header->eth;
    uint32_t f, ra[2], *macp, rctl = core->mac[RCTL];
    uint16_t queues = 0;
    uint16_t oversized = 0;
    size_t vlan_num = 0;
    PTP2 ptp2;
    bool lpe;
    uint16_t rlpml;
    int i;

    memset(rss_info, 0, sizeof(E1000E_RSSInfo));
    *ts = false;

    if (external_tx) {
        *external_tx = true;
    }

    if (core->mac[CTRL_EXT] & BIT(26)) {
        if (be16_to_cpu(ehdr->h_proto) == core->mac[VET] >> 16 &&
            be16_to_cpu(l2_header->vlan[0].h_proto) == (core->mac[VET] & 0xffff)) {
            vlan_num = 2;
        }
    } else {
        if (be16_to_cpu(ehdr->h_proto) == (core->mac[VET] & 0xffff)) {
            vlan_num = 1;
        }
    }

    lpe = !!(core->mac[RCTL] & E1000_RCTL_LPE);
    rlpml = core->mac[RLPML];
    if (!(core->mac[RCTL] & E1000_RCTL_SBP) &&
        igb_rx_is_oversized(core, ehdr, size, vlan_num, lpe, rlpml)) {
        trace_e1000x_rx_oversized(size);
        return queues;
    }

    for (*etqf = 0; *etqf < 8; (*etqf)++) {
        if ((core->mac[ETQF0 + *etqf] & E1000_ETQF_FILTER_ENABLE) &&
            be16_to_cpu(ehdr->h_proto) == (core->mac[ETQF0 + *etqf] & E1000_ETQF_ETYPE_MASK)) {
            if ((core->mac[ETQF0 + *etqf] & E1000_ETQF_1588) &&
                (core->mac[TSYNCRXCTL] & E1000_TSYNCRXCTL_ENABLED) &&
                !(core->mac[TSYNCRXCTL] & E1000_TSYNCRXCTL_VALID) &&
                iov_to_buf(iov, iovcnt, iov_ofs + ETH_HLEN, &ptp2, sizeof(ptp2)) >= sizeof(ptp2) &&
                (ptp2.version_ptp & 15) == 2 &&
                ptp2.message_id_transport_specific == ((core->mac[TSYNCRXCFG] >> 8) & 255)) {
                e1000x_timestamp(core->mac, core->timadj, RXSTMPL, RXSTMPH);
                *ts = true;
                core->mac[TSYNCRXCTL] |= E1000_TSYNCRXCTL_VALID;
                core->mac[RXSATRL] = le32_to_cpu(ptp2.source_uuid_lo);
                core->mac[RXSATRH] = le16_to_cpu(ptp2.source_uuid_hi) |
                                     (le16_to_cpu(ptp2.sequence_id) << 16);
            }
            break;
        }
    }

    if (vlan_num &&
        !e1000x_rx_vlan_filter(core->mac, l2_header->vlan + vlan_num - 1)) {
        return queues;
    }

    if (core->mac[MRQC] & 1) {
        if (is_broadcast_ether_addr(ehdr->h_dest)) {
            for (i = 0; i < IGB_NUM_VM_POOLS; i++) {
                if (core->mac[VMOLR0 + i] & E1000_VMOLR_BAM) {
                    queues |= BIT(i);
                }
            }
        } else {
            for (macp = core->mac + RA; macp < core->mac + RA + 32; macp += 2) {
                if (!(macp[1] & E1000_RAH_AV)) {
                    continue;
                }
                ra[0] = cpu_to_le32(macp[0]);
                ra[1] = cpu_to_le32(macp[1]);
                if (!memcmp(ehdr->h_dest, (uint8_t *)ra, ETH_ALEN)) {
                    queues |= (macp[1] & E1000_RAH_POOL_MASK) / E1000_RAH_POOL_1;
                }
            }

            for (macp = core->mac + RA2; macp < core->mac + RA2 + 16; macp += 2) {
                if (!(macp[1] & E1000_RAH_AV)) {
                    continue;
                }
                ra[0] = cpu_to_le32(macp[0]);
                ra[1] = cpu_to_le32(macp[1]);
                if (!memcmp(ehdr->h_dest, (uint8_t *)ra, ETH_ALEN)) {
                    queues |= (macp[1] & E1000_RAH_POOL_MASK) / E1000_RAH_POOL_1;
                }
            }

            if (!queues) {
                macp = core->mac + (is_multicast_ether_addr(ehdr->h_dest) ? MTA : UTA);

                f = ta_shift[(rctl >> E1000_RCTL_MO_SHIFT) & 3];
                f = (((ehdr->h_dest[5] << 8) | ehdr->h_dest[4]) >> f) & 0xfff;
                if (macp[f >> 5] & (1 << (f & 0x1f))) {
                    for (i = 0; i < IGB_NUM_VM_POOLS; i++) {
                        if (core->mac[VMOLR0 + i] & E1000_VMOLR_ROMPE) {
                            queues |= BIT(i);
                        }
                    }
                }
            } else if (is_unicast_ether_addr(ehdr->h_dest) && external_tx) {
                *external_tx = false;
            }
        }

        if (e1000x_vlan_rx_filter_enabled(core->mac)) {
            uint16_t mask = 0;

            if (vlan_num) {
                uint16_t vid = be16_to_cpu(l2_header->vlan[vlan_num - 1].h_tci) & VLAN_VID_MASK;

                for (i = 0; i < E1000_VLVF_ARRAY_SIZE; i++) {
                    if ((core->mac[VLVF0 + i] & E1000_VLVF_VLANID_MASK) == vid &&
                        (core->mac[VLVF0 + i] & E1000_VLVF_VLANID_ENABLE)) {
                        uint32_t poolsel = core->mac[VLVF0 + i] & E1000_VLVF_POOLSEL_MASK;
                        mask |= poolsel >> E1000_VLVF_POOLSEL_SHIFT;
                    }
                }
            } else {
                for (i = 0; i < IGB_NUM_VM_POOLS; i++) {
                    if (core->mac[VMOLR0 + i] & E1000_VMOLR_AUPE) {
                        mask |= BIT(i);
                    }
                }
            }

            queues &= mask;
        }

        if (is_unicast_ether_addr(ehdr->h_dest) && !queues && !external_tx &&
            !(core->mac[VT_CTL] & E1000_VT_CTL_DISABLE_DEF_POOL)) {
            uint32_t def_pl = core->mac[VT_CTL] & E1000_VT_CTL_DEFAULT_POOL_MASK;
            queues = BIT(def_pl >> E1000_VT_CTL_DEFAULT_POOL_SHIFT);
        }

        queues &= core->mac[VFRE];
        if (queues) {
            for (i = 0; i < IGB_NUM_VM_POOLS; i++) {
                lpe = !!(core->mac[VMOLR0 + i] & E1000_VMOLR_LPE);
                rlpml = core->mac[VMOLR0 + i] & E1000_VMOLR_RLPML_MASK;
                if ((queues & BIT(i)) &&
                    igb_rx_is_oversized(core, ehdr, size, vlan_num,
                                        lpe, rlpml)) {
                    oversized |= BIT(i);
                }
            }
            /* 8.19.37 increment ROC if packet is oversized for all queues */
            if (oversized == queues) {
                trace_e1000x_rx_oversized(size);
                e1000x_inc_reg_if_not_full(core->mac, ROC);
            }
            queues &= ~oversized;
        }

        if (queues) {
            igb_rss_parse_packet(core, core->rx_pkt,
                                 external_tx != NULL, rss_info);
            /* Sec 8.26.1: PQn = VFn + VQn*8 */
            if (rss_info->queue & 1) {
                for (i = 0; i < IGB_NUM_VM_POOLS; i++) {
                    if ((queues & BIT(i)) &&
                        (core->mac[VMOLR0 + i] & E1000_VMOLR_RSSE)) {
                        queues |= BIT(i + IGB_NUM_VM_POOLS);
                        queues &= ~BIT(i);
                    }
                }
            }
        }
    } else {
        bool accepted = e1000x_rx_group_filter(core->mac, ehdr);
        if (!accepted) {
            for (macp = core->mac + RA2; macp < core->mac + RA2 + 16; macp += 2) {
                if (!(macp[1] & E1000_RAH_AV)) {
                    continue;
                }
                ra[0] = cpu_to_le32(macp[0]);
                ra[1] = cpu_to_le32(macp[1]);
                if (!memcmp(ehdr->h_dest, (uint8_t *)ra, ETH_ALEN)) {
                    trace_e1000x_rx_flt_ucast_match((int)(macp - core->mac - RA2) / 2,
                                                    MAC_ARG(ehdr->h_dest));

                    accepted = true;
                    break;
                }
            }
        }

        if (accepted) {
            igb_rss_parse_packet(core, core->rx_pkt, false, rss_info);
            queues = BIT(rss_info->queue);
        }
    }

    return queues;
}

static inline void
igb_read_lgcy_rx_descr(IGBCore *core, struct e1000_rx_desc *desc,
                       hwaddr *buff_addr)
{
    *buff_addr = le64_to_cpu(desc->buffer_addr);
}

static inline void
igb_read_adv_rx_descr(IGBCore *core, union e1000_adv_rx_desc *desc,
                      hwaddr *buff_addr)
{
    *buff_addr = le64_to_cpu(desc->read.pkt_addr);
}

static inline void
igb_read_rx_descr(IGBCore *core, union e1000_rx_desc_union *desc,
                  hwaddr *buff_addr)
{
    if (igb_rx_use_legacy_descriptor(core)) {
        igb_read_lgcy_rx_descr(core, &desc->legacy, buff_addr);
    } else {
        igb_read_adv_rx_descr(core, &desc->adv, buff_addr);
    }
}

static void
igb_verify_csum_in_sw(IGBCore *core,
                      struct NetRxPkt *pkt,
                      uint32_t *status_flags,
                      EthL4HdrProto l4hdr_proto)
{
    bool csum_valid;
    uint32_t csum_error;

    if (igb_rx_l3_cso_enabled(core)) {
        if (!net_rx_pkt_validate_l3_csum(pkt, &csum_valid)) {
            trace_e1000e_rx_metadata_l3_csum_validation_failed();
        } else {
            csum_error = csum_valid ? 0 : E1000_RXDEXT_STATERR_IPE;
            *status_flags |= E1000_RXD_STAT_IPCS | csum_error;
        }
    } else {
        trace_e1000e_rx_metadata_l3_cso_disabled();
    }

    if (!igb_rx_l4_cso_enabled(core)) {
        trace_e1000e_rx_metadata_l4_cso_disabled();
        return;
    }

    if (!net_rx_pkt_validate_l4_csum(pkt, &csum_valid)) {
        trace_e1000e_rx_metadata_l4_csum_validation_failed();
        return;
    }

    csum_error = csum_valid ? 0 : E1000_RXDEXT_STATERR_TCPE;
    *status_flags |= E1000_RXD_STAT_TCPCS | csum_error;

    if (l4hdr_proto == ETH_L4_HDR_PROTO_UDP) {
        *status_flags |= E1000_RXD_STAT_UDPCS;
    }
}

static void
igb_build_rx_metadata(IGBCore *core,
                      struct NetRxPkt *pkt,
                      bool is_eop,
                      const E1000E_RSSInfo *rss_info, uint16_t etqf, bool ts,
                      uint16_t *pkt_info, uint16_t *hdr_info,
                      uint32_t *rss,
                      uint32_t *status_flags,
                      uint16_t *ip_id,
                      uint16_t *vlan_tag)
{
    struct virtio_net_hdr *vhdr;
    bool hasip4, hasip6, csum_valid;
    EthL4HdrProto l4hdr_proto;

    *status_flags = E1000_RXD_STAT_DD;

    /* No additional metadata needed for non-EOP descriptors */
    /* TODO: EOP apply only to status so don't skip whole function. */
    if (!is_eop) {
        goto func_exit;
    }

    *status_flags |= E1000_RXD_STAT_EOP;

    net_rx_pkt_get_protocols(pkt, &hasip4, &hasip6, &l4hdr_proto);
    trace_e1000e_rx_metadata_protocols(hasip4, hasip6, l4hdr_proto);

    /* VLAN state */
    if (net_rx_pkt_is_vlan_stripped(pkt)) {
        *status_flags |= E1000_RXD_STAT_VP;
        *vlan_tag = cpu_to_le16(net_rx_pkt_get_vlan_tag(pkt));
        trace_e1000e_rx_metadata_vlan(*vlan_tag);
    }

    /* Packet parsing results */
    if ((core->mac[RXCSUM] & E1000_RXCSUM_PCSD) != 0) {
        if (rss_info->enabled) {
            *rss = cpu_to_le32(rss_info->hash);
            trace_igb_rx_metadata_rss(*rss);
        }
    } else if (hasip4) {
            *status_flags |= E1000_RXD_STAT_IPIDV;
            *ip_id = cpu_to_le16(net_rx_pkt_get_ip_id(pkt));
            trace_e1000e_rx_metadata_ip_id(*ip_id);
    }

    if (l4hdr_proto == ETH_L4_HDR_PROTO_TCP && net_rx_pkt_is_tcp_ack(pkt)) {
        *status_flags |= E1000_RXD_STAT_ACK;
        trace_e1000e_rx_metadata_ack();
    }

    if (pkt_info) {
        *pkt_info = rss_info->enabled ? rss_info->type : 0;

        if (etqf < 8) {
            *pkt_info |= (BIT(11) | etqf) << 4;
        } else {
            if (hasip4) {
                *pkt_info |= E1000_ADVRXD_PKT_IP4;
            }

            if (hasip6) {
                *pkt_info |= E1000_ADVRXD_PKT_IP6;
            }

            switch (l4hdr_proto) {
            case ETH_L4_HDR_PROTO_TCP:
                *pkt_info |= E1000_ADVRXD_PKT_TCP;
                break;

            case ETH_L4_HDR_PROTO_UDP:
                *pkt_info |= E1000_ADVRXD_PKT_UDP;
                break;

            case ETH_L4_HDR_PROTO_SCTP:
                *pkt_info |= E1000_ADVRXD_PKT_SCTP;
                break;

            default:
                break;
            }
        }
    }

    if (hdr_info) {
        *hdr_info = 0;
    }

    if (ts) {
        *status_flags |= BIT(16);
    }

    /* RX CSO information */
    if (hasip6 && (core->mac[RFCTL] & E1000_RFCTL_IPV6_XSUM_DIS)) {
        trace_e1000e_rx_metadata_ipv6_sum_disabled();
        goto func_exit;
    }

    vhdr = net_rx_pkt_get_vhdr(pkt);

    if (!(vhdr->flags & VIRTIO_NET_HDR_F_DATA_VALID) &&
        !(vhdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM)) {
        trace_e1000e_rx_metadata_virthdr_no_csum_info();
        igb_verify_csum_in_sw(core, pkt, status_flags, l4hdr_proto);
        goto func_exit;
    }

    if (igb_rx_l3_cso_enabled(core)) {
        *status_flags |= hasip4 ? E1000_RXD_STAT_IPCS : 0;
    } else {
        trace_e1000e_rx_metadata_l3_cso_disabled();
    }

    if (igb_rx_l4_cso_enabled(core)) {
        switch (l4hdr_proto) {
        case ETH_L4_HDR_PROTO_SCTP:
            if (!net_rx_pkt_validate_l4_csum(pkt, &csum_valid)) {
                trace_e1000e_rx_metadata_l4_csum_validation_failed();
                goto func_exit;
            }
            if (!csum_valid) {
                *status_flags |= E1000_RXDEXT_STATERR_TCPE;
            }
            /* fall through */
        case ETH_L4_HDR_PROTO_TCP:
            *status_flags |= E1000_RXD_STAT_TCPCS;
            break;

        case ETH_L4_HDR_PROTO_UDP:
            *status_flags |= E1000_RXD_STAT_TCPCS | E1000_RXD_STAT_UDPCS;
            break;

        default:
            break;
        }
    } else {
        trace_e1000e_rx_metadata_l4_cso_disabled();
    }

func_exit:
    trace_e1000e_rx_metadata_status_flags(*status_flags);
    *status_flags = cpu_to_le32(*status_flags);
}

static inline void
igb_write_lgcy_rx_descr(IGBCore *core, struct e1000_rx_desc *desc,
                        struct NetRxPkt *pkt,
                        const E1000E_RSSInfo *rss_info, uint16_t etqf, bool ts,
                        uint16_t length)
{
    uint32_t status_flags, rss;
    uint16_t ip_id;

    assert(!rss_info->enabled);
    desc->length = cpu_to_le16(length);
    desc->csum = 0;

    igb_build_rx_metadata(core, pkt, pkt != NULL,
                          rss_info, etqf, ts,
                          NULL, NULL, &rss,
                          &status_flags, &ip_id,
                          &desc->special);
    desc->errors = (uint8_t) (le32_to_cpu(status_flags) >> 24);
    desc->status = (uint8_t) le32_to_cpu(status_flags);
}

static inline void
igb_write_adv_rx_descr(IGBCore *core, union e1000_adv_rx_desc *desc,
                       struct NetRxPkt *pkt,
                       const E1000E_RSSInfo *rss_info, uint16_t etqf, bool ts,
                       uint16_t length)
{
    memset(&desc->wb, 0, sizeof(desc->wb));

    desc->wb.upper.length = cpu_to_le16(length);

    igb_build_rx_metadata(core, pkt, pkt != NULL,
                          rss_info, etqf, ts,
                          &desc->wb.lower.lo_dword.pkt_info,
                          &desc->wb.lower.lo_dword.hdr_info,
                          &desc->wb.lower.hi_dword.rss,
                          &desc->wb.upper.status_error,
                          &desc->wb.lower.hi_dword.csum_ip.ip_id,
                          &desc->wb.upper.vlan);
}

static inline void
igb_write_rx_descr(IGBCore *core, union e1000_rx_desc_union *desc,
                   struct NetRxPkt *pkt, const E1000E_RSSInfo *rss_info,
                   uint16_t etqf, bool ts, uint16_t length)
{
    if (igb_rx_use_legacy_descriptor(core)) {
        igb_write_lgcy_rx_descr(core, &desc->legacy, pkt, rss_info,
                                etqf, ts, length);
    } else {
        igb_write_adv_rx_descr(core, &desc->adv, pkt, rss_info,
                               etqf, ts, length);
    }
}

static inline void
igb_pci_dma_write_rx_desc(IGBCore *core, PCIDevice *dev, dma_addr_t addr,
                          union e1000_rx_desc_union *desc, dma_addr_t len)
{
    if (igb_rx_use_legacy_descriptor(core)) {
        struct e1000_rx_desc *d = &desc->legacy;
        size_t offset = offsetof(struct e1000_rx_desc, status);
        uint8_t status = d->status;

        d->status &= ~E1000_RXD_STAT_DD;
        pci_dma_write(dev, addr, desc, len);

        if (status & E1000_RXD_STAT_DD) {
            d->status = status;
            pci_dma_write(dev, addr + offset, &status, sizeof(status));
        }
    } else {
        union e1000_adv_rx_desc *d = &desc->adv;
        size_t offset =
            offsetof(union e1000_adv_rx_desc, wb.upper.status_error);
        uint32_t status = d->wb.upper.status_error;

        d->wb.upper.status_error &= ~E1000_RXD_STAT_DD;
        pci_dma_write(dev, addr, desc, len);

        if (status & E1000_RXD_STAT_DD) {
            d->wb.upper.status_error = status;
            pci_dma_write(dev, addr + offset, &status, sizeof(status));
        }
    }
}

static void
igb_write_to_rx_buffers(IGBCore *core,
                        PCIDevice *d,
                        hwaddr ba,
                        uint16_t *written,
                        const char *data,
                        dma_addr_t data_len)
{
    trace_igb_rx_desc_buff_write(ba, *written, data, data_len);
    pci_dma_write(d, ba + *written, data, data_len);
    *written += data_len;
}

static void
igb_update_rx_stats(IGBCore *core, const E1000E_RingInfo *rxi,
                    size_t pkt_size, size_t pkt_fcs_size)
{
    eth_pkt_types_e pkt_type = net_rx_pkt_get_packet_type(core->rx_pkt);
    e1000x_update_rx_total_stats(core->mac, pkt_type, pkt_size, pkt_fcs_size);

    if (core->mac[MRQC] & 1) {
        uint16_t pool = rxi->idx % IGB_NUM_VM_POOLS;

        core->mac[PVFGORC0 + (pool * 64)] += pkt_size + 4;
        core->mac[PVFGPRC0 + (pool * 64)]++;
        if (pkt_type == ETH_PKT_MCAST) {
            core->mac[PVFMPRC0 + (pool * 64)]++;
        }
    }
}

static inline bool
igb_rx_descr_threshold_hit(IGBCore *core, const E1000E_RingInfo *rxi)
{
    return igb_ring_free_descr_num(core, rxi) ==
           ((core->mac[E1000_SRRCTL(rxi->idx) >> 2] >> 20) & 31) * 16;
}

static void
igb_write_packet_to_guest(IGBCore *core, struct NetRxPkt *pkt,
                          const E1000E_RxRing *rxr,
                          const E1000E_RSSInfo *rss_info,
                          uint16_t etqf, bool ts)
{
    PCIDevice *d;
    dma_addr_t base;
    union e1000_rx_desc_union desc;
    size_t desc_size;
    size_t desc_offset = 0;
    size_t iov_ofs = 0;

    struct iovec *iov = net_rx_pkt_get_iovec(pkt);
    size_t size = net_rx_pkt_get_total_len(pkt);
    size_t total_size = size + e1000x_fcs_len(core->mac);
    const E1000E_RingInfo *rxi = rxr->i;
    size_t bufsize = igb_rxbufsize(core, rxi);

    d = pcie_sriov_get_vf_at_index(core->owner, rxi->idx % 8);
    if (!d) {
        d = core->owner;
    }

    do {
        hwaddr ba;
        uint16_t written = 0;
        bool is_last = false;

        desc_size = total_size - desc_offset;

        if (desc_size > bufsize) {
            desc_size = bufsize;
        }

        if (igb_ring_empty(core, rxi)) {
            return;
        }

        base = igb_ring_head_descr(core, rxi);

        pci_dma_read(d, base, &desc, core->rx_desc_len);

        trace_e1000e_rx_descr(rxi->idx, base, core->rx_desc_len);

        igb_read_rx_descr(core, &desc, &ba);

        if (ba) {
            if (desc_offset < size) {
                static const uint32_t fcs_pad;
                size_t iov_copy;
                size_t copy_size = size - desc_offset;
                if (copy_size > bufsize) {
                    copy_size = bufsize;
                }

                /* Copy packet payload */
                while (copy_size) {
                    iov_copy = MIN(copy_size, iov->iov_len - iov_ofs);

                    igb_write_to_rx_buffers(core, d, ba, &written,
                                            iov->iov_base + iov_ofs, iov_copy);

                    copy_size -= iov_copy;
                    iov_ofs += iov_copy;
                    if (iov_ofs == iov->iov_len) {
                        iov++;
                        iov_ofs = 0;
                    }
                }

                if (desc_offset + desc_size >= total_size) {
                    /* Simulate FCS checksum presence in the last descriptor */
                    igb_write_to_rx_buffers(core, d, ba, &written,
                          (const char *) &fcs_pad, e1000x_fcs_len(core->mac));
                }
            }
        } else { /* as per intel docs; skip descriptors with null buf addr */
            trace_e1000e_rx_null_descriptor();
        }
        desc_offset += desc_size;
        if (desc_offset >= total_size) {
            is_last = true;
        }

        igb_write_rx_descr(core, &desc, is_last ? core->rx_pkt : NULL,
                           rss_info, etqf, ts, written);
        igb_pci_dma_write_rx_desc(core, d, base, &desc, core->rx_desc_len);

        igb_ring_advance(core, rxi, core->rx_desc_len / E1000_MIN_RX_DESC_LEN);

    } while (desc_offset < total_size);

    igb_update_rx_stats(core, rxi, size, total_size);
}

static bool
igb_rx_strip_vlan(IGBCore *core, const E1000E_RingInfo *rxi)
{
    if (core->mac[MRQC] & 1) {
        uint16_t pool = rxi->idx % IGB_NUM_VM_POOLS;
        /* Sec 7.10.3.8: CTRL.VME is ignored, only VMOLR/RPLOLR is used */
        return (net_rx_pkt_get_packet_type(core->rx_pkt) == ETH_PKT_MCAST) ?
                core->mac[RPLOLR] & E1000_RPLOLR_STRVLAN :
                core->mac[VMOLR0 + pool] & E1000_VMOLR_STRVLAN;
    }

    return e1000x_vlan_enabled(core->mac);
}

static inline void
igb_rx_fix_l4_csum(IGBCore *core, struct NetRxPkt *pkt)
{
    struct virtio_net_hdr *vhdr = net_rx_pkt_get_vhdr(pkt);

    if (vhdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
        net_rx_pkt_fix_l4_csum(pkt);
    }
}

ssize_t
igb_receive_iov(IGBCore *core, const struct iovec *iov, int iovcnt)
{
    return igb_receive_internal(core, iov, iovcnt, core->has_vnet, NULL);
}

static ssize_t
igb_receive_internal(IGBCore *core, const struct iovec *iov, int iovcnt,
                     bool has_vnet, bool *external_tx)
{
    uint16_t queues = 0;
    uint32_t causes = 0;
    uint32_t ecauses = 0;
    union {
        L2Header l2_header;
        uint8_t octets[ETH_ZLEN];
    } buf;
    struct iovec min_iov;
    size_t size, orig_size;
    size_t iov_ofs = 0;
    E1000E_RxRing rxr;
    E1000E_RSSInfo rss_info;
    uint16_t etqf;
    bool ts;
    size_t total_size;
    int strip_vlan_index;
    int i;

    trace_e1000e_rx_receive_iov(iovcnt);

    if (external_tx) {
        *external_tx = true;
    }

    if (!e1000x_hw_rx_enabled(core->mac)) {
        return -1;
    }

    /* Pull virtio header in */
    if (has_vnet) {
        net_rx_pkt_set_vhdr_iovec(core->rx_pkt, iov, iovcnt);
        iov_ofs = sizeof(struct virtio_net_hdr);
    } else {
        net_rx_pkt_unset_vhdr(core->rx_pkt);
    }

    orig_size = iov_size(iov, iovcnt);
    size = orig_size - iov_ofs;

    /* Pad to minimum Ethernet frame length */
    if (size < sizeof(buf)) {
        iov_to_buf(iov, iovcnt, iov_ofs, &buf, size);
        memset(&buf.octets[size], 0, sizeof(buf) - size);
        e1000x_inc_reg_if_not_full(core->mac, RUC);
        min_iov.iov_base = &buf;
        min_iov.iov_len = size = sizeof(buf);
        iovcnt = 1;
        iov = &min_iov;
        iov_ofs = 0;
    } else {
        iov_to_buf(iov, iovcnt, iov_ofs, &buf, sizeof(buf.l2_header));
    }

    net_rx_pkt_set_packet_type(core->rx_pkt,
                               get_eth_packet_type(&buf.l2_header.eth));
    net_rx_pkt_set_protocols(core->rx_pkt, iov, iovcnt, iov_ofs);

    queues = igb_receive_assign(core, iov, iovcnt, iov_ofs,
                                &buf.l2_header, size,
                                &rss_info, &etqf, &ts, external_tx);
    if (!queues) {
        trace_e1000e_rx_flt_dropped();
        return orig_size;
    }

    for (i = 0; i < IGB_NUM_QUEUES; i++) {
        if (!(queues & BIT(i)) ||
            !(core->mac[RXDCTL0 + (i * 16)] & E1000_RXDCTL_QUEUE_ENABLE)) {
            continue;
        }

        igb_rx_ring_init(core, &rxr, i);

        if (!igb_rx_strip_vlan(core, rxr.i)) {
            strip_vlan_index = -1;
        } else if (core->mac[CTRL_EXT] & BIT(26)) {
            strip_vlan_index = 1;
        } else {
            strip_vlan_index = 0;
        }

        net_rx_pkt_attach_iovec_ex(core->rx_pkt, iov, iovcnt, iov_ofs,
                                   strip_vlan_index,
                                   core->mac[VET] & 0xffff,
                                   core->mac[VET] >> 16);

        total_size = net_rx_pkt_get_total_len(core->rx_pkt) +
            e1000x_fcs_len(core->mac);

        if (!igb_has_rxbufs(core, rxr.i, total_size)) {
            causes |= E1000_ICS_RXO;
            trace_e1000e_rx_not_written_to_guest(rxr.i->idx);
            continue;
        }

        causes |= E1000_ICR_RXDW;

        igb_rx_fix_l4_csum(core, core->rx_pkt);
        igb_write_packet_to_guest(core, core->rx_pkt, &rxr, &rss_info, etqf, ts);

        /* Check if receive descriptor minimum threshold hit */
        if (igb_rx_descr_threshold_hit(core, rxr.i)) {
            causes |= E1000_ICS_RXDMT0;
        }

        ecauses |= igb_rx_wb_eic(core, rxr.i->idx);

        trace_e1000e_rx_written_to_guest(rxr.i->idx);
    }

    trace_e1000e_rx_interrupt_set(causes);
    igb_raise_interrupts(core, EICR, ecauses);
    igb_raise_interrupts(core, ICR, causes);

    return orig_size;
}

static inline bool
igb_have_autoneg(IGBCore *core)
{
    return core->phy[MII_BMCR] & MII_BMCR_AUTOEN;
}

static void igb_update_flowctl_status(IGBCore *core)
{
    if (igb_have_autoneg(core) && core->phy[MII_BMSR] & MII_BMSR_AN_COMP) {
        trace_e1000e_link_autoneg_flowctl(true);
        core->mac[CTRL] |= E1000_CTRL_TFCE | E1000_CTRL_RFCE;
    } else {
        trace_e1000e_link_autoneg_flowctl(false);
    }
}

static inline void
igb_link_down(IGBCore *core)
{
    e1000x_update_regs_on_link_down(core->mac, core->phy);
    igb_update_flowctl_status(core);
}

static inline void
igb_set_phy_ctrl(IGBCore *core, uint16_t val)
{
    /* bits 0-5 reserved; MII_BMCR_[ANRESTART,RESET] are self clearing */
    core->phy[MII_BMCR] = val & ~(0x3f | MII_BMCR_RESET | MII_BMCR_ANRESTART);

    if ((val & MII_BMCR_ANRESTART) && igb_have_autoneg(core)) {
        e1000x_restart_autoneg(core->mac, core->phy, core->autoneg_timer);
    }
}

void igb_core_set_link_status(IGBCore *core)
{
    NetClientState *nc = qemu_get_queue(core->owner_nic);
    uint32_t old_status = core->mac[STATUS];

    trace_e1000e_link_status_changed(nc->link_down ? false : true);

    if (nc->link_down) {
        e1000x_update_regs_on_link_down(core->mac, core->phy);
    } else {
        if (igb_have_autoneg(core) &&
            !(core->phy[MII_BMSR] & MII_BMSR_AN_COMP)) {
            e1000x_restart_autoneg(core->mac, core->phy,
                                   core->autoneg_timer);
        } else {
            e1000x_update_regs_on_link_up(core->mac, core->phy);
            igb_start_recv(core);
        }
    }

    if (core->mac[STATUS] != old_status) {
        igb_raise_interrupts(core, ICR, E1000_ICR_LSC);
    }
}

static void
igb_set_ctrl(IGBCore *core, int index, uint32_t val)
{
    trace_e1000e_core_ctrl_write(index, val);

    /* RST is self clearing */
    core->mac[CTRL] = val & ~E1000_CTRL_RST;
    core->mac[CTRL_DUP] = core->mac[CTRL];

    trace_e1000e_link_set_params(
        !!(val & E1000_CTRL_ASDE),
        (val & E1000_CTRL_SPD_SEL) >> E1000_CTRL_SPD_SHIFT,
        !!(val & E1000_CTRL_FRCSPD),
        !!(val & E1000_CTRL_FRCDPX),
        !!(val & E1000_CTRL_RFCE),
        !!(val & E1000_CTRL_TFCE));

    if (val & E1000_CTRL_RST) {
        trace_e1000e_core_ctrl_sw_reset();
        igb_reset(core, true);
    }

    if (val & E1000_CTRL_PHY_RST) {
        trace_e1000e_core_ctrl_phy_reset();
        core->mac[STATUS] |= E1000_STATUS_PHYRA;
    }
}

static void
igb_set_rfctl(IGBCore *core, int index, uint32_t val)
{
    trace_e1000e_rx_set_rfctl(val);

    if (!(val & E1000_RFCTL_ISCSI_DIS)) {
        trace_e1000e_wrn_iscsi_filtering_not_supported();
    }

    if (!(val & E1000_RFCTL_NFSW_DIS)) {
        trace_e1000e_wrn_nfsw_filtering_not_supported();
    }

    if (!(val & E1000_RFCTL_NFSR_DIS)) {
        trace_e1000e_wrn_nfsr_filtering_not_supported();
    }

    core->mac[RFCTL] = val;
}

static void
igb_calc_rxdesclen(IGBCore *core)
{
    if (igb_rx_use_legacy_descriptor(core)) {
        core->rx_desc_len = sizeof(struct e1000_rx_desc);
    } else {
        core->rx_desc_len = sizeof(union e1000_adv_rx_desc);
    }
    trace_e1000e_rx_desc_len(core->rx_desc_len);
}

static void
igb_set_rx_control(IGBCore *core, int index, uint32_t val)
{
    core->mac[RCTL] = val;
    trace_e1000e_rx_set_rctl(core->mac[RCTL]);

    if (val & E1000_RCTL_DTYP_MASK) {
        qemu_log_mask(LOG_GUEST_ERROR,
                      "igb: RCTL.DTYP must be zero for compatibility");
    }

    if (val & E1000_RCTL_EN) {
        igb_calc_rxdesclen(core);
        igb_start_recv(core);
    }
}

static inline bool
igb_postpone_interrupt(IGBIntrDelayTimer *timer)
{
    if (timer->running) {
        trace_e1000e_irq_postponed_by_xitr(timer->delay_reg << 2);

        return true;
    }

    if (timer->core->mac[timer->delay_reg] != 0) {
        igb_intrmgr_rearm_timer(timer);
    }

    return false;
}

static inline bool
igb_eitr_should_postpone(IGBCore *core, int idx)
{
    return igb_postpone_interrupt(&core->eitr[idx]);
}

static void igb_send_msix(IGBCore *core, uint32_t causes)
{
    int vector;

    for (vector = 0; vector < IGB_INTR_NUM; ++vector) {
        if ((causes & BIT(vector)) && !igb_eitr_should_postpone(core, vector)) {

            trace_e1000e_irq_msix_notify_vec(vector);
            igb_msix_notify(core, vector);
        }
    }
}

static inline void
igb_fix_icr_asserted(IGBCore *core)
{
    core->mac[ICR] &= ~E1000_ICR_ASSERTED;
    if (core->mac[ICR]) {
        core->mac[ICR] |= E1000_ICR_ASSERTED;
    }

    trace_e1000e_irq_fix_icr_asserted(core->mac[ICR]);
}

static void igb_raise_interrupts(IGBCore *core, size_t index, uint32_t causes)
{
    uint32_t old_causes = core->mac[ICR] & core->mac[IMS];
    uint32_t old_ecauses = core->mac[EICR] & core->mac[EIMS];
    uint32_t raised_causes;
    uint32_t raised_ecauses;
    uint32_t int_alloc;

    trace_e1000e_irq_set(index << 2,
                         core->mac[index], core->mac[index] | causes);

    core->mac[index] |= causes;

    if (core->mac[GPIE] & E1000_GPIE_MSIX_MODE) {
        raised_causes = core->mac[ICR] & core->mac[IMS] & ~old_causes;

        if (raised_causes & E1000_ICR_DRSTA) {
            int_alloc = core->mac[IVAR_MISC] & 0xff;
            if (int_alloc & E1000_IVAR_VALID) {

                core->mac[EICR] |= BIT(int_alloc & 0x1f);
            }
        }
        /* Check if other bits (excluding the TCP Timer) are enabled. */
        if (raised_causes & ~E1000_ICR_DRSTA) {
            int_alloc = (core->mac[IVAR_MISC] >> 8) & 0xff;
            if (int_alloc & E1000_IVAR_VALID) {
                core->mac[EICR] |= BIT(int_alloc & 0x1f);
            }
        }

        raised_ecauses = core->mac[EICR] & core->mac[EIMS] & ~old_ecauses;
        if (!raised_ecauses) {
            return;
        }

        igb_send_msix(core, raised_ecauses);
    } else {
        igb_fix_icr_asserted(core);

        raised_causes = core->mac[ICR] & core->mac[IMS] & ~old_causes;
        if (!raised_causes) {
            return;
        }

        core->mac[EICR] |= (raised_causes & E1000_ICR_DRSTA) | E1000_EICR_OTHER;

        if (msix_enabled(core->owner)) {
            trace_e1000e_irq_msix_notify_vec(0);
            msix_notify(core->owner, 0);
        } else if (msi_enabled(core->owner)) {
            trace_e1000e_irq_msi_notify(raised_causes);
            msi_notify(core->owner, 0);
        } else {
            igb_raise_legacy_irq(core);
        }
    }
}

static void igb_lower_interrupts(IGBCore *core, size_t index, uint32_t causes)
{
    trace_e1000e_irq_clear(index << 2,
                           core->mac[index], core->mac[index] & ~causes);

    core->mac[index] &= ~causes;

    trace_e1000e_irq_pending_interrupts(core->mac[ICR] & core->mac[IMS],
                                        core->mac[ICR], core->mac[IMS]);

    if (!(core->mac[ICR] & core->mac[IMS]) &&
        !(core->mac[GPIE] & E1000_GPIE_MSIX_MODE)) {
        core->mac[EICR] &= ~E1000_EICR_OTHER;

        if (!msix_enabled(core->owner) && !msi_enabled(core->owner)) {
            igb_lower_legacy_irq(core);
        }
    }
}

static void igb_set_eics(IGBCore *core, int index, uint32_t val)
{
    bool msix = !!(core->mac[GPIE] & E1000_GPIE_MSIX_MODE);
    uint32_t mask = msix ? E1000_EICR_MSIX_MASK : E1000_EICR_LEGACY_MASK;

    trace_igb_irq_write_eics(val, msix);
    igb_raise_interrupts(core, EICR, val & mask);
}

static void igb_set_eims(IGBCore *core, int index, uint32_t val)
{
    bool msix = !!(core->mac[GPIE] & E1000_GPIE_MSIX_MODE);
    uint32_t mask = msix ? E1000_EICR_MSIX_MASK : E1000_EICR_LEGACY_MASK;

    trace_igb_irq_write_eims(val, msix);
    igb_raise_interrupts(core, EIMS, val & mask);
}

static void mailbox_interrupt_to_vf(IGBCore *core, uint16_t vfn)
{
    uint32_t ent = core->mac[VTIVAR_MISC + vfn];
    uint32_t causes;

    if ((ent & E1000_IVAR_VALID)) {
        causes = (ent & 0x3) << (22 - vfn * IGBVF_MSIX_VEC_NUM);
        igb_raise_interrupts(core, EICR, causes);
    }
}

static void mailbox_interrupt_to_pf(IGBCore *core)
{
    igb_raise_interrupts(core, ICR, E1000_ICR_VMMB);
}

static void igb_set_pfmailbox(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = index - P2VMAILBOX0;

    trace_igb_set_pfmailbox(vfn, val);

    if (val & E1000_P2VMAILBOX_STS) {
        core->mac[V2PMAILBOX0 + vfn] |= E1000_V2PMAILBOX_PFSTS;
        mailbox_interrupt_to_vf(core, vfn);
    }

    if (val & E1000_P2VMAILBOX_ACK) {
        core->mac[V2PMAILBOX0 + vfn] |= E1000_V2PMAILBOX_PFACK;
        mailbox_interrupt_to_vf(core, vfn);
    }

    /* Buffer Taken by PF (can be set only if the VFU is cleared). */
    if (val & E1000_P2VMAILBOX_PFU) {
        if (!(core->mac[index] & E1000_P2VMAILBOX_VFU)) {
            core->mac[index] |= E1000_P2VMAILBOX_PFU;
            core->mac[V2PMAILBOX0 + vfn] |= E1000_V2PMAILBOX_PFU;
        }
    } else {
        core->mac[index] &= ~E1000_P2VMAILBOX_PFU;
        core->mac[V2PMAILBOX0 + vfn] &= ~E1000_V2PMAILBOX_PFU;
    }

    if (val & E1000_P2VMAILBOX_RVFU) {
        core->mac[V2PMAILBOX0 + vfn] &= ~E1000_V2PMAILBOX_VFU;
        core->mac[MBVFICR] &= ~((E1000_MBVFICR_VFACK_VF1 << vfn) |
                                (E1000_MBVFICR_VFREQ_VF1 << vfn));
    }
}

static void igb_set_vfmailbox(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = index - V2PMAILBOX0;

    trace_igb_set_vfmailbox(vfn, val);

    if (val & E1000_V2PMAILBOX_REQ) {
        core->mac[MBVFICR] |= E1000_MBVFICR_VFREQ_VF1 << vfn;
        mailbox_interrupt_to_pf(core);
    }

    if (val & E1000_V2PMAILBOX_ACK) {
        core->mac[MBVFICR] |= E1000_MBVFICR_VFACK_VF1 << vfn;
        mailbox_interrupt_to_pf(core);
    }

    /* Buffer Taken by VF (can be set only if the PFU is cleared). */
    if (val & E1000_V2PMAILBOX_VFU) {
        if (!(core->mac[index] & E1000_V2PMAILBOX_PFU)) {
            core->mac[index] |= E1000_V2PMAILBOX_VFU;
            core->mac[P2VMAILBOX0 + vfn] |= E1000_P2VMAILBOX_VFU;
        }
    } else {
        core->mac[index] &= ~E1000_V2PMAILBOX_VFU;
        core->mac[P2VMAILBOX0 + vfn] &= ~E1000_P2VMAILBOX_VFU;
    }
}

static void igb_vf_reset(IGBCore *core, uint16_t vfn)
{
    uint16_t qn0 = vfn;
    uint16_t qn1 = vfn + IGB_NUM_VM_POOLS;

    /* disable Rx and Tx for the VF*/
    core->mac[RXDCTL0 + (qn0 * 16)] &= ~E1000_RXDCTL_QUEUE_ENABLE;
    core->mac[RXDCTL0 + (qn1 * 16)] &= ~E1000_RXDCTL_QUEUE_ENABLE;
    core->mac[TXDCTL0 + (qn0 * 16)] &= ~E1000_TXDCTL_QUEUE_ENABLE;
    core->mac[TXDCTL0 + (qn1 * 16)] &= ~E1000_TXDCTL_QUEUE_ENABLE;
    core->mac[VFRE] &= ~BIT(vfn);
    core->mac[VFTE] &= ~BIT(vfn);
    /* indicate VF reset to PF */
    core->mac[VFLRE] |= BIT(vfn);
    /* VFLRE and mailbox use the same interrupt cause */
    mailbox_interrupt_to_pf(core);
}

static void igb_w1c(IGBCore *core, int index, uint32_t val)
{
    core->mac[index] &= ~val;
}

static void igb_set_eimc(IGBCore *core, int index, uint32_t val)
{
    bool msix = !!(core->mac[GPIE] & E1000_GPIE_MSIX_MODE);
    uint32_t mask = msix ? E1000_EICR_MSIX_MASK : E1000_EICR_LEGACY_MASK;

    trace_igb_irq_write_eimc(val, msix);

    /* Interrupts are disabled via a write to EIMC and reflected in EIMS. */
    igb_lower_interrupts(core, EIMS, val & mask);
}

static void igb_set_eiac(IGBCore *core, int index, uint32_t val)
{
    bool msix = !!(core->mac[GPIE] & E1000_GPIE_MSIX_MODE);

    if (msix) {
        trace_igb_irq_write_eiac(val);

        /*
         * TODO: When using IOV, the bits that correspond to MSI-X vectors
         * that are assigned to a VF are read-only.
         */
        core->mac[EIAC] |= (val & E1000_EICR_MSIX_MASK);
    }
}

static void igb_set_eiam(IGBCore *core, int index, uint32_t val)
{
    bool msix = !!(core->mac[GPIE] & E1000_GPIE_MSIX_MODE);

    /*
     * TODO: When using IOV, the bits that correspond to MSI-X vectors that
     * are assigned to a VF are read-only.
     */
    core->mac[EIAM] |=
        ~(val & (msix ? E1000_EICR_MSIX_MASK : E1000_EICR_LEGACY_MASK));

    trace_igb_irq_write_eiam(val, msix);
}

static void igb_set_eicr(IGBCore *core, int index, uint32_t val)
{
    bool msix = !!(core->mac[GPIE] & E1000_GPIE_MSIX_MODE);

    /*
     * TODO: In IOV mode, only bit zero of this vector is available for the PF
     * function.
     */
    uint32_t mask = msix ? E1000_EICR_MSIX_MASK : E1000_EICR_LEGACY_MASK;

    trace_igb_irq_write_eicr(val, msix);
    igb_lower_interrupts(core, EICR, val & mask);
}

static void igb_set_vtctrl(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn;

    if (val & E1000_CTRL_RST) {
        vfn = (index - PVTCTRL0) / 0x40;
        igb_vf_reset(core, vfn);
    }
}

static void igb_set_vteics(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = (index - PVTEICS0) / 0x40;

    core->mac[index] = val;
    igb_set_eics(core, EICS, (val & 0x7) << (22 - vfn * IGBVF_MSIX_VEC_NUM));
}

static void igb_set_vteims(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = (index - PVTEIMS0) / 0x40;

    core->mac[index] = val;
    igb_set_eims(core, EIMS, (val & 0x7) << (22 - vfn * IGBVF_MSIX_VEC_NUM));
}

static void igb_set_vteimc(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = (index - PVTEIMC0) / 0x40;

    core->mac[index] = val;
    igb_set_eimc(core, EIMC, (val & 0x7) << (22 - vfn * IGBVF_MSIX_VEC_NUM));
}

static void igb_set_vteiac(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = (index - PVTEIAC0) / 0x40;

    core->mac[index] = val;
    igb_set_eiac(core, EIAC, (val & 0x7) << (22 - vfn * IGBVF_MSIX_VEC_NUM));
}

static void igb_set_vteiam(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = (index - PVTEIAM0) / 0x40;

    core->mac[index] = val;
    igb_set_eiam(core, EIAM, (val & 0x7) << (22 - vfn * IGBVF_MSIX_VEC_NUM));
}

static void igb_set_vteicr(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = (index - PVTEICR0) / 0x40;

    core->mac[index] = val;
    igb_set_eicr(core, EICR, (val & 0x7) << (22 - vfn * IGBVF_MSIX_VEC_NUM));
}

static void igb_set_vtivar(IGBCore *core, int index, uint32_t val)
{
    uint16_t vfn = (index - VTIVAR);
    uint16_t qn = vfn;
    uint8_t ent;
    int n;

    core->mac[index] = val;

    /* Get assigned vector associated with queue Rx#0. */
    if ((val & E1000_IVAR_VALID)) {
        n = igb_ivar_entry_rx(qn);
        ent = E1000_IVAR_VALID | (24 - vfn * IGBVF_MSIX_VEC_NUM - (2 - (val & 0x7)));
        core->mac[IVAR0 + n / 4] |= ent << 8 * (n % 4);
    }

    /* Get assigned vector associated with queue Tx#0 */
    ent = val >> 8;
    if ((ent & E1000_IVAR_VALID)) {
        n = igb_ivar_entry_tx(qn);
        ent = E1000_IVAR_VALID | (24 - vfn * IGBVF_MSIX_VEC_NUM - (2 - (ent & 0x7)));
        core->mac[IVAR0 + n / 4] |= ent << 8 * (n % 4);
    }

    /*
     * Ignoring assigned vectors associated with queues Rx#1 and Tx#1 for now.
     */
}

static inline void
igb_autoneg_timer(void *opaque)
{
    IGBCore *core = opaque;
    if (!qemu_get_queue(core->owner_nic)->link_down) {
        e1000x_update_regs_on_autoneg_done(core->mac, core->phy);
        igb_start_recv(core);

        igb_update_flowctl_status(core);
        /* signal link status change to the guest */
        igb_raise_interrupts(core, ICR, E1000_ICR_LSC);
    }
}

static inline uint16_t
igb_get_reg_index_with_offset(const uint16_t *mac_reg_access, hwaddr addr)
{
    uint16_t index = (addr & 0x1ffff) >> 2;
    return index + (mac_reg_access[index] & 0xfffe);
}

static const char igb_phy_regcap[MAX_PHY_REG_ADDRESS + 1] = {
    [MII_BMCR]                   = PHY_RW,
    [MII_BMSR]                   = PHY_R,
    [MII_PHYID1]                 = PHY_R,
    [MII_PHYID2]                 = PHY_R,
    [MII_ANAR]                   = PHY_RW,
    [MII_ANLPAR]                 = PHY_R,
    [MII_ANER]                   = PHY_R,
    [MII_ANNP]                   = PHY_RW,
    [MII_ANLPRNP]                = PHY_R,
    [MII_CTRL1000]               = PHY_RW,
    [MII_STAT1000]               = PHY_R,
    [MII_EXTSTAT]                = PHY_R,

    [IGP01E1000_PHY_PORT_CONFIG] = PHY_RW,
    [IGP01E1000_PHY_PORT_STATUS] = PHY_R,
    [IGP01E1000_PHY_PORT_CTRL]   = PHY_RW,
    [IGP01E1000_PHY_LINK_HEALTH] = PHY_R,
    [IGP02E1000_PHY_POWER_MGMT]  = PHY_RW,
    [IGP01E1000_PHY_PAGE_SELECT] = PHY_W
};

static void
igb_phy_reg_write(IGBCore *core, uint32_t addr, uint16_t data)
{
    assert(addr <= MAX_PHY_REG_ADDRESS);

    if (addr == MII_BMCR) {
        igb_set_phy_ctrl(core, data);
    } else {
        core->phy[addr] = data;
    }
}

static void
igb_set_mdic(IGBCore *core, int index, uint32_t val)
{
    uint32_t data = val & E1000_MDIC_DATA_MASK;
    uint32_t addr = ((val & E1000_MDIC_REG_MASK) >> E1000_MDIC_REG_SHIFT);

    if ((val & E1000_MDIC_PHY_MASK) >> E1000_MDIC_PHY_SHIFT != 1) { /* phy # */
        val = core->mac[MDIC] | E1000_MDIC_ERROR;
    } else if (val & E1000_MDIC_OP_READ) {
        if (!(igb_phy_regcap[addr] & PHY_R)) {
            trace_igb_core_mdic_read_unhandled(addr);
            val |= E1000_MDIC_ERROR;
        } else {
            val = (val ^ data) | core->phy[addr];
            trace_igb_core_mdic_read(addr, val);
        }
    } else if (val & E1000_MDIC_OP_WRITE) {
        if (!(igb_phy_regcap[addr] & PHY_W)) {
            trace_igb_core_mdic_write_unhandled(addr);
            val |= E1000_MDIC_ERROR;
        } else {
            trace_igb_core_mdic_write(addr, data);
            igb_phy_reg_write(core, addr, data);
        }
    }
    core->mac[MDIC] = val | E1000_MDIC_READY;

    if (val & E1000_MDIC_INT_EN) {
        igb_raise_interrupts(core, ICR, E1000_ICR_MDAC);
    }
}

static void
igb_set_rdt(IGBCore *core, int index, uint32_t val)
{
    core->mac[index] = val & 0xffff;
    trace_e1000e_rx_set_rdt(igb_mq_queue_idx(RDT0, index), val);
    igb_start_recv(core);
}

static void
igb_set_status(IGBCore *core, int index, uint32_t val)
{
    if ((val & E1000_STATUS_PHYRA) == 0) {
        core->mac[index] &= ~E1000_STATUS_PHYRA;
    }
}

static void
igb_set_ctrlext(IGBCore *core, int index, uint32_t val)
{
    trace_igb_link_set_ext_params(!!(val & E1000_CTRL_EXT_ASDCHK),
                                  !!(val & E1000_CTRL_EXT_SPD_BYPS),
                                  !!(val & E1000_CTRL_EXT_PFRSTD));

    /* Zero self-clearing bits */
    val &= ~(E1000_CTRL_EXT_ASDCHK | E1000_CTRL_EXT_EE_RST);
    core->mac[CTRL_EXT] = val;

    if (core->mac[CTRL_EXT] & E1000_CTRL_EXT_PFRSTD) {
        for (int vfn = 0; vfn < IGB_MAX_VF_FUNCTIONS; vfn++) {
            core->mac[V2PMAILBOX0 + vfn] &= ~E1000_V2PMAILBOX_RSTI;
            core->mac[V2PMAILBOX0 + vfn] |= E1000_V2PMAILBOX_RSTD;
        }
    }
}

static void
igb_set_pbaclr(IGBCore *core, int index, uint32_t val)
{
    int i;

    core->mac[PBACLR] = val & E1000_PBACLR_VALID_MASK;

    if (!msix_enabled(core->owner)) {
        return;
    }

    for (i = 0; i < IGB_INTR_NUM; i++) {
        if (core->mac[PBACLR] & BIT(i)) {
            msix_clr_pending(core->owner, i);
        }
    }
}

static void
igb_set_fcrth(IGBCore *core, int index, uint32_t val)
{
    core->mac[FCRTH] = val & 0xFFF8;
}

static void
igb_set_fcrtl(IGBCore *core, int index, uint32_t val)
{
    core->mac[FCRTL] = val & 0x8000FFF8;
}

#define IGB_LOW_BITS_SET_FUNC(num)                             \
    static void                                                \
    igb_set_##num##bit(IGBCore *core, int index, uint32_t val) \
    {                                                          \
        core->mac[index] = val & (BIT(num) - 1);               \
    }

IGB_LOW_BITS_SET_FUNC(4)
IGB_LOW_BITS_SET_FUNC(13)
IGB_LOW_BITS_SET_FUNC(16)

static void
igb_set_dlen(IGBCore *core, int index, uint32_t val)
{
    core->mac[index] = val & 0xffff0;
}

static void
igb_set_dbal(IGBCore *core, int index, uint32_t val)
{
    core->mac[index] = val & E1000_XDBAL_MASK;
}

static void
igb_set_tdt(IGBCore *core, int index, uint32_t val)
{
    IGB_TxRing txr;
    int qn = igb_mq_queue_idx(TDT0, index);

    core->mac[index] = val & 0xffff;

    igb_tx_ring_init(core, &txr, qn);
    igb_start_xmit(core, &txr);
}

static void
igb_set_ics(IGBCore *core, int index, uint32_t val)
{
    trace_e1000e_irq_write_ics(val);
    igb_raise_interrupts(core, ICR, val);
}

static void
igb_set_imc(IGBCore *core, int index, uint32_t val)
{
    trace_e1000e_irq_ims_clear_set_imc(val);
    igb_lower_interrupts(core, IMS, val);
}

static void
igb_set_ims(IGBCore *core, int index, uint32_t val)
{
    igb_raise_interrupts(core, IMS, val & 0x77D4FBFD);
}

static void igb_nsicr(IGBCore *core)
{
    /*
     * If GPIE.NSICR = 0, then the clear of IMS will occur only if at
     * least one bit is set in the IMS and there is a true interrupt as
     * reflected in ICR.INTA.
     */
    if ((core->mac[GPIE] & E1000_GPIE_NSICR) ||
        (core->mac[IMS] && (core->mac[ICR] & E1000_ICR_INT_ASSERTED))) {
        igb_lower_interrupts(core, IMS, core->mac[IAM]);
    }
}

static void igb_set_icr(IGBCore *core, int index, uint32_t val)
{
    igb_nsicr(core);
    igb_lower_interrupts(core, ICR, val);
}

static uint32_t
igb_mac_readreg(IGBCore *core, int index)
{
    return core->mac[index];
}

static uint32_t
igb_mac_ics_read(IGBCore *core, int index)
{
    trace_e1000e_irq_read_ics(core->mac[ICS]);
    return core->mac[ICS];
}

static uint32_t
igb_mac_ims_read(IGBCore *core, int index)
{
    trace_e1000e_irq_read_ims(core->mac[IMS]);
    return core->mac[IMS];
}

static uint32_t
igb_mac_swsm_read(IGBCore *core, int index)
{
    uint32_t val = core->mac[SWSM];
    core->mac[SWSM] = val | E1000_SWSM_SMBI;
    return val;
}

static uint32_t
igb_mac_eitr_read(IGBCore *core, int index)
{
    return core->eitr_guest_value[index - EITR0];
}

static uint32_t igb_mac_vfmailbox_read(IGBCore *core, int index)
{
    uint32_t val = core->mac[index];

    core->mac[index] &= ~(E1000_V2PMAILBOX_PFSTS | E1000_V2PMAILBOX_PFACK |
                          E1000_V2PMAILBOX_RSTD);

    return val;
}

static uint32_t
igb_mac_icr_read(IGBCore *core, int index)
{
    uint32_t ret = core->mac[ICR];

    if (core->mac[GPIE] & E1000_GPIE_NSICR) {
        trace_igb_irq_icr_clear_gpie_nsicr();
        igb_lower_interrupts(core, ICR, 0xffffffff);
    } else if (core->mac[IMS] == 0) {
        trace_e1000e_irq_icr_clear_zero_ims();
        igb_lower_interrupts(core, ICR, 0xffffffff);
    } else if (core->mac[ICR] & E1000_ICR_INT_ASSERTED) {
        igb_lower_interrupts(core, ICR, 0xffffffff);
    } else if (!msix_enabled(core->owner)) {
        trace_e1000e_irq_icr_clear_nonmsix_icr_read();
        igb_lower_interrupts(core, ICR, 0xffffffff);
    }

    igb_nsicr(core);
    return ret;
}

static uint32_t
igb_mac_read_clr4(IGBCore *core, int index)
{
    uint32_t ret = core->mac[index];

    core->mac[index] = 0;
    return ret;
}

static uint32_t
igb_mac_read_clr8(IGBCore *core, int index)
{
    uint32_t ret = core->mac[index];

    core->mac[index] = 0;
    core->mac[index - 1] = 0;
    return ret;
}

static uint32_t
igb_get_ctrl(IGBCore *core, int index)
{
    uint32_t val = core->mac[CTRL];

    trace_e1000e_link_read_params(
        !!(val & E1000_CTRL_ASDE),
        (val & E1000_CTRL_SPD_SEL) >> E1000_CTRL_SPD_SHIFT,
        !!(val & E1000_CTRL_FRCSPD),
        !!(val & E1000_CTRL_FRCDPX),
        !!(val & E1000_CTRL_RFCE),
        !!(val & E1000_CTRL_TFCE));

    return val;
}

static uint32_t igb_get_status(IGBCore *core, int index)
{
    uint32_t res = core->mac[STATUS];
    uint16_t num_vfs = pcie_sriov_num_vfs(core->owner);

    if (core->mac[CTRL] & E1000_CTRL_FRCDPX) {
        res |= (core->mac[CTRL] & E1000_CTRL_FD) ? E1000_STATUS_FD : 0;
    } else {
        res |= E1000_STATUS_FD;
    }

    if ((core->mac[CTRL] & E1000_CTRL_FRCSPD) ||
        (core->mac[CTRL_EXT] & E1000_CTRL_EXT_SPD_BYPS)) {
        switch (core->mac[CTRL] & E1000_CTRL_SPD_SEL) {
        case E1000_CTRL_SPD_10:
            res |= E1000_STATUS_SPEED_10;
            break;
        case E1000_CTRL_SPD_100:
            res |= E1000_STATUS_SPEED_100;
            break;
        case E1000_CTRL_SPD_1000:
        default:
            res |= E1000_STATUS_SPEED_1000;
            break;
        }
    } else {
        res |= E1000_STATUS_SPEED_1000;
    }

    if (num_vfs) {
        res |= num_vfs << E1000_STATUS_NUM_VFS_SHIFT;
        res |= E1000_STATUS_IOV_MODE;
    }

    if (!(core->mac[CTRL] & E1000_CTRL_GIO_MASTER_DISABLE)) {
        res |= E1000_STATUS_GIO_MASTER_ENABLE;
    }

    return res;
}

static void
igb_mac_writereg(IGBCore *core, int index, uint32_t val)
{
    core->mac[index] = val;
}

static void
igb_mac_setmacaddr(IGBCore *core, int index, uint32_t val)
{
    uint32_t macaddr[2];

    core->mac[index] = val;

    macaddr[0] = cpu_to_le32(core->mac[RA]);
    macaddr[1] = cpu_to_le32(core->mac[RA + 1]);
    qemu_format_nic_info_str(qemu_get_queue(core->owner_nic),
        (uint8_t *) macaddr);

    trace_e1000e_mac_set_sw(MAC_ARG(macaddr));
}

static void
igb_set_eecd(IGBCore *core, int index, uint32_t val)
{
    static const uint32_t ro_bits = E1000_EECD_PRES          |
                                    E1000_EECD_AUTO_RD       |
                                    E1000_EECD_SIZE_EX_MASK;

    core->mac[EECD] = (core->mac[EECD] & ro_bits) | (val & ~ro_bits);
}

static void
igb_set_eerd(IGBCore *core, int index, uint32_t val)
{
    uint32_t addr = (val >> E1000_EERW_ADDR_SHIFT) & E1000_EERW_ADDR_MASK;
    uint32_t flags = 0;
    uint32_t data = 0;

    if ((addr < IGB_EEPROM_SIZE) && (val & E1000_EERW_START)) {
        data = core->eeprom[addr];
        flags = E1000_EERW_DONE;
    }

    core->mac[EERD] = flags                           |
                      (addr << E1000_EERW_ADDR_SHIFT) |
                      (data << E1000_EERW_DATA_SHIFT);
}

static void
igb_set_eitr(IGBCore *core, int index, uint32_t val)
{
    uint32_t eitr_num = index - EITR0;

    trace_igb_irq_eitr_set(eitr_num, val);

    core->eitr_guest_value[eitr_num] = val & ~E1000_EITR_CNT_IGNR;
    core->mac[index] = val & 0x7FFE;
}

static void
igb_update_rx_offloads(IGBCore *core)
{
    int cso_state = igb_rx_l4_cso_enabled(core);

    trace_e1000e_rx_set_cso(cso_state);

    if (core->has_vnet) {
        qemu_set_offload(qemu_get_queue(core->owner_nic)->peer,
                         cso_state, 0, 0, 0, 0);
    }
}

static void
igb_set_rxcsum(IGBCore *core, int index, uint32_t val)
{
    core->mac[RXCSUM] = val;
    igb_update_rx_offloads(core);
}

static void
igb_set_gcr(IGBCore *core, int index, uint32_t val)
{
    uint32_t ro_bits = core->mac[GCR] & E1000_GCR_RO_BITS;
    core->mac[GCR] = (val & ~E1000_GCR_RO_BITS) | ro_bits;
}

static uint32_t igb_get_systiml(IGBCore *core, int index)
{
    e1000x_timestamp(core->mac, core->timadj, SYSTIML, SYSTIMH);
    return core->mac[SYSTIML];
}

static uint32_t igb_get_rxsatrh(IGBCore *core, int index)
{
    core->mac[TSYNCRXCTL] &= ~E1000_TSYNCRXCTL_VALID;
    return core->mac[RXSATRH];
}

static uint32_t igb_get_txstmph(IGBCore *core, int index)
{
    core->mac[TSYNCTXCTL] &= ~E1000_TSYNCTXCTL_VALID;
    return core->mac[TXSTMPH];
}

static void igb_set_timinca(IGBCore *core, int index, uint32_t val)
{
    e1000x_set_timinca(core->mac, &core->timadj, val);
}

static void igb_set_timadjh(IGBCore *core, int index, uint32_t val)
{
    core->mac[TIMADJH] = val;
    core->timadj += core->mac[TIMADJL] | ((int64_t)core->mac[TIMADJH] << 32);
}

#define igb_getreg(x)    [x] = igb_mac_readreg
typedef uint32_t (*readops)(IGBCore *, int);
static const readops igb_macreg_readops[] = {
    igb_getreg(WUFC),
    igb_getreg(MANC),
    igb_getreg(TOTL),
    igb_getreg(RDT0),
    igb_getreg(RDT1),
    igb_getreg(RDT2),
    igb_getreg(RDT3),
    igb_getreg(RDT4),
    igb_getreg(RDT5),
    igb_getreg(RDT6),
    igb_getreg(RDT7),
    igb_getreg(RDT8),
    igb_getreg(RDT9),
    igb_getreg(RDT10),
    igb_getreg(RDT11),
    igb_getreg(RDT12),
    igb_getreg(RDT13),
    igb_getreg(RDT14),
    igb_getreg(RDT15),
    igb_getreg(RDBAH0),
    igb_getreg(RDBAH1),
    igb_getreg(RDBAH2),
    igb_getreg(RDBAH3),
    igb_getreg(RDBAH4),
    igb_getreg(RDBAH5),
    igb_getreg(RDBAH6),
    igb_getreg(RDBAH7),
    igb_getreg(RDBAH8),
    igb_getreg(RDBAH9),
    igb_getreg(RDBAH10),
    igb_getreg(RDBAH11),
    igb_getreg(RDBAH12),
    igb_getreg(RDBAH13),
    igb_getreg(RDBAH14),
    igb_getreg(RDBAH15),
    igb_getreg(TDBAL0),
    igb_getreg(TDBAL1),
    igb_getreg(TDBAL2),
    igb_getreg(TDBAL3),
    igb_getreg(TDBAL4),
    igb_getreg(TDBAL5),
    igb_getreg(TDBAL6),
    igb_getreg(TDBAL7),
    igb_getreg(TDBAL8),
    igb_getreg(TDBAL9),
    igb_getreg(TDBAL10),
    igb_getreg(TDBAL11),
    igb_getreg(TDBAL12),
    igb_getreg(TDBAL13),
    igb_getreg(TDBAL14),
    igb_getreg(TDBAL15),
    igb_getreg(RDLEN0),
    igb_getreg(RDLEN1),
    igb_getreg(RDLEN2),
    igb_getreg(RDLEN3),
    igb_getreg(RDLEN4),
    igb_getreg(RDLEN5),
    igb_getreg(RDLEN6),
    igb_getreg(RDLEN7),
    igb_getreg(RDLEN8),
    igb_getreg(RDLEN9),
    igb_getreg(RDLEN10),
    igb_getreg(RDLEN11),
    igb_getreg(RDLEN12),
    igb_getreg(RDLEN13),
    igb_getreg(RDLEN14),
    igb_getreg(RDLEN15),
    igb_getreg(SRRCTL0),
    igb_getreg(SRRCTL1),
    igb_getreg(SRRCTL2),
    igb_getreg(SRRCTL3),
    igb_getreg(SRRCTL4),
    igb_getreg(SRRCTL5),
    igb_getreg(SRRCTL6),
    igb_getreg(SRRCTL7),
    igb_getreg(SRRCTL8),
    igb_getreg(SRRCTL9),
    igb_getreg(SRRCTL10),
    igb_getreg(SRRCTL11),
    igb_getreg(SRRCTL12),
    igb_getreg(SRRCTL13),
    igb_getreg(SRRCTL14),
    igb_getreg(SRRCTL15),
    igb_getreg(LATECOL),
    igb_getreg(XONTXC),
    igb_getreg(TDFH),
    igb_getreg(TDFT),
    igb_getreg(TDFHS),
    igb_getreg(TDFTS),
    igb_getreg(TDFPC),
    igb_getreg(WUS),
    igb_getreg(RDFH),
    igb_getreg(RDFT),
    igb_getreg(RDFHS),
    igb_getreg(RDFTS),
    igb_getreg(RDFPC),
    igb_getreg(GORCL),
    igb_getreg(MGTPRC),
    igb_getreg(EERD),
    igb_getreg(EIAC),
    igb_getreg(MANC2H),
    igb_getreg(RXCSUM),
    igb_getreg(GSCL_3),
    igb_getreg(GSCN_2),
    igb_getreg(FCAH),
    igb_getreg(FCRTH),
    igb_getreg(FLOP),
    igb_getreg(RXSTMPH),
    igb_getreg(TXSTMPL),
    igb_getreg(TIMADJL),
    igb_getreg(RDH0),
    igb_getreg(RDH1),
    igb_getreg(RDH2),
    igb_getreg(RDH3),
    igb_getreg(RDH4),
    igb_getreg(RDH5),
    igb_getreg(RDH6),
    igb_getreg(RDH7),
    igb_getreg(RDH8),
    igb_getreg(RDH9),
    igb_getreg(RDH10),
    igb_getreg(RDH11),
    igb_getreg(RDH12),
    igb_getreg(RDH13),
    igb_getreg(RDH14),
    igb_getreg(RDH15),
    igb_getreg(TDT0),
    igb_getreg(TDT1),
    igb_getreg(TDT2),
    igb_getreg(TDT3),
    igb_getreg(TDT4),
    igb_getreg(TDT5),
    igb_getreg(TDT6),
    igb_getreg(TDT7),
    igb_getreg(TDT8),
    igb_getreg(TDT9),
    igb_getreg(TDT10),
    igb_getreg(TDT11),
    igb_getreg(TDT12),
    igb_getreg(TDT13),
    igb_getreg(TDT14),
    igb_getreg(TDT15),
    igb_getreg(TNCRS),
    igb_getreg(RJC),
    igb_getreg(IAM),
    igb_getreg(GSCL_2),
    igb_getreg(TIPG),
    igb_getreg(FLMNGCTL),
    igb_getreg(FLMNGCNT),
    igb_getreg(TSYNCTXCTL),
    igb_getreg(EEMNGDATA),
    igb_getreg(CTRL_EXT),
    igb_getreg(SYSTIMH),
    igb_getreg(EEMNGCTL),
    igb_getreg(FLMNGDATA),
    igb_getreg(TSYNCRXCTL),
    igb_getreg(LEDCTL),
    igb_getreg(TCTL),
    igb_getreg(TCTL_EXT),
    igb_getreg(DTXCTL),
    igb_getreg(RXPBS),
    igb_getreg(TDH0),
    igb_getreg(TDH1),
    igb_getreg(TDH2),
    igb_getreg(TDH3),
    igb_getreg(TDH4),
    igb_getreg(TDH5),
    igb_getreg(TDH6),
    igb_getreg(TDH7),
    igb_getreg(TDH8),
    igb_getreg(TDH9),
    igb_getreg(TDH10),
    igb_getreg(TDH11),
    igb_getreg(TDH12),
    igb_getreg(TDH13),
    igb_getreg(TDH14),
    igb_getreg(TDH15),
    igb_getreg(ECOL),
    igb_getreg(DC),
    igb_getreg(RLEC),
    igb_getreg(XOFFTXC),
    igb_getreg(RFC),
    igb_getreg(RNBC),
    igb_getreg(MGTPTC),
    igb_getreg(TIMINCA),
    igb_getreg(FACTPS),
    igb_getreg(GSCL_1),
    igb_getreg(GSCN_0),
    igb_getreg(PBACLR),
    igb_getreg(FCTTV),
    igb_getreg(RXSATRL),
    igb_getreg(TORL),
    igb_getreg(TDLEN0),
    igb_getreg(TDLEN1),
    igb_getreg(TDLEN2),