// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
/* Copyright (C) 2015-2018 Netronome Systems, Inc. */

/*
 * nfp_net_common.c
 * Netronome network device driver: Common functions between PF and VF
 * Authors: Jakub Kicinski <jakub.kicinski@netronome.com>
 *          Jason McMullan <jason.mcmullan@netronome.com>
 *          Rolf Neugebauer <rolf.neugebauer@netronome.com>
 *          Brad Petrus <brad.petrus@netronome.com>
 *          Chris Telfer <chris.telfer@netronome.com>
 */

#include <linux/bitfield.h>
#include <linux/bpf.h>
#include <linux/bpf_trace.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/fs.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/interrupt.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/mm.h>
#include <linux/overflow.h>
#include <linux/page_ref.h>
#include <linux/pci.h>
#include <linux/pci_regs.h>
#include <linux/msi.h>
#include <linux/ethtool.h>
#include <linux/log2.h>
#include <linux/if_vlan.h>
#include <linux/random.h>
#include <linux/vmalloc.h>
#include <linux/ktime.h>

#include <net/tls.h>
#include <net/vxlan.h>

#include "nfpcore/nfp_nsp.h"
#include "ccm.h"
#include "nfp_app.h"
#include "nfp_net_ctrl.h"
#include "nfp_net.h"
#include "nfp_net_sriov.h"
#include "nfp_port.h"
#include "crypto/crypto.h"
#include "crypto/fw.h"

/**
 * nfp_net_get_fw_version() - Read and parse the FW version
 * @fw_ver:     Output fw_version structure to read to
 * @ctrl_bar:   Mapped address of the control BAR
 */
void nfp_net_get_fw_version(struct nfp_net_fw_version *fw_ver,
                            void __iomem *ctrl_bar)
{
        u32 reg;

        reg = readl(ctrl_bar + NFP_NET_CFG_VERSION);
        put_unaligned_le32(reg, fw_ver);
}

static dma_addr_t nfp_net_dma_map_rx(struct nfp_net_dp *dp, void *frag)
{
        return dma_map_single_attrs(dp->dev, frag + NFP_NET_RX_BUF_HEADROOM,
                                    dp->fl_bufsz - NFP_NET_RX_BUF_NON_DATA,
                                    dp->rx_dma_dir, DMA_ATTR_SKIP_CPU_SYNC);
}

static void
nfp_net_dma_sync_dev_rx(const struct nfp_net_dp *dp, dma_addr_t dma_addr)
{
        dma_sync_single_for_device(dp->dev, dma_addr,
                                   dp->fl_bufsz - NFP_NET_RX_BUF_NON_DATA,
                                   dp->rx_dma_dir);
}

static void nfp_net_dma_unmap_rx(struct nfp_net_dp *dp, dma_addr_t dma_addr)
{
        dma_unmap_single_attrs(dp->dev, dma_addr,
                               dp->fl_bufsz - NFP_NET_RX_BUF_NON_DATA,
                               dp->rx_dma_dir, DMA_ATTR_SKIP_CPU_SYNC);
}

static void nfp_net_dma_sync_cpu_rx(struct nfp_net_dp *dp, dma_addr_t dma_addr,
                                    unsigned int len)
{
        dma_sync_single_for_cpu(dp->dev, dma_addr - NFP_NET_RX_BUF_HEADROOM,
                                len, dp->rx_dma_dir);
}

/* Firmware reconfig
 *
 * Firmware reconfig may take a while so we have two versions of it -
 * synchronous and asynchronous (posted).  All synchronous callers are holding
 * RTNL so we don't have to worry about serializing them.
 */
static void nfp_net_reconfig_start(struct nfp_net *nn, u32 update)
{
        nn_writel(nn, NFP_NET_CFG_UPDATE, update);
        /* ensure update is written before pinging HW */
        nn_pci_flush(nn);
        nfp_qcp_wr_ptr_add(nn->qcp_cfg, 1);
        nn->reconfig_in_progress_update = update;
}

/* Pass 0 as update to run posted reconfigs. */
static void nfp_net_reconfig_start_async(struct nfp_net *nn, u32 update)
{
        update |= nn->reconfig_posted;
        nn->reconfig_posted = 0;

        nfp_net_reconfig_start(nn, update);

        nn->reconfig_timer_active = true;
        mod_timer(&nn->reconfig_timer, jiffies + NFP_NET_POLL_TIMEOUT * HZ);
}

static bool nfp_net_reconfig_check_done(struct nfp_net *nn, bool last_check)
{
        u32 reg;

        reg = nn_readl(nn, NFP_NET_CFG_UPDATE);
        if (reg == 0)
                return true;
        if (reg & NFP_NET_CFG_UPDATE_ERR) {
                nn_err(nn, "Reconfig error (status: 0x%08x update: 0x%08x ctrl: 0x%08x)\n",
                       reg, nn->reconfig_in_progress_update,
                       nn_readl(nn, NFP_NET_CFG_CTRL));
                return true;
        } else if (last_check) {
                nn_err(nn, "Reconfig timeout (status: 0x%08x update: 0x%08x ctrl: 0x%08x)\n",
                       reg, nn->reconfig_in_progress_update,
                       nn_readl(nn, NFP_NET_CFG_CTRL));
                return true;
        }

        return false;
}

static bool __nfp_net_reconfig_wait(struct nfp_net *nn, unsigned long deadline)
{
        bool timed_out = false;
        int i;

        /* Poll update field, waiting for NFP to ack the config.
         * Do an opportunistic wait-busy loop, afterward sleep.
         */
        for (i = 0; i < 50; i++) {
                if (nfp_net_reconfig_check_done(nn, false))
                        return false;
                udelay(4);
        }

        while (!nfp_net_reconfig_check_done(nn, timed_out)) {
                usleep_range(250, 500);
                timed_out = time_is_before_eq_jiffies(deadline);
        }

        return timed_out;
}

static int nfp_net_reconfig_wait(struct nfp_net *nn, unsigned long deadline)
{
        if (__nfp_net_reconfig_wait(nn, deadline))
                return -EIO;

        if (nn_readl(nn, NFP_NET_CFG_UPDATE) & NFP_NET_CFG_UPDATE_ERR)
                return -EIO;

        return 0;
}

static void nfp_net_reconfig_timer(struct timer_list *t)
{
        struct nfp_net *nn = from_timer(nn, t, reconfig_timer);

        spin_lock_bh(&nn->reconfig_lock);

        nn->reconfig_timer_active = false;

        /* If sync caller is present it will take over from us */
        if (nn->reconfig_sync_present)
                goto done;

        /* Read reconfig status and report errors */
        nfp_net_reconfig_check_done(nn, true);

        if (nn->reconfig_posted)
                nfp_net_reconfig_start_async(nn, 0);
done:
        spin_unlock_bh(&nn->reconfig_lock);
}

/**
 * nfp_net_reconfig_post() - Post async reconfig request
 * @nn:      NFP Net device to reconfigure
 * @update:  The value for the update field in the BAR config
 *
 * Record FW reconfiguration request.  Reconfiguration will be kicked off
 * whenever reconfiguration machinery is idle.  Multiple requests can be
 * merged together!
 */
static void nfp_net_reconfig_post(struct nfp_net *nn, u32 update)
{
        spin_lock_bh(&nn->reconfig_lock);

        /* Sync caller will kick off async reconf when it's done, just post */
        if (nn->reconfig_sync_present) {
                nn->reconfig_posted |= update;
                goto done;
        }

        /* Opportunistically check if the previous command is done */
        if (!nn->reconfig_timer_active ||
            nfp_net_reconfig_check_done(nn, false))
                nfp_net_reconfig_start_async(nn, update);
        else
                nn->reconfig_posted |= update;
done:
        spin_unlock_bh(&nn->reconfig_lock);
}

static void nfp_net_reconfig_sync_enter(struct nfp_net *nn)
{
        bool cancelled_timer = false;
        u32 pre_posted_requests;

        spin_lock_bh(&nn->reconfig_lock);

        WARN_ON(nn->reconfig_sync_present);
        nn->reconfig_sync_present = true;

        if (nn->reconfig_timer_active) {
                nn->reconfig_timer_active = false;
                cancelled_timer = true;
        }
        pre_posted_requests = nn->reconfig_posted;
        nn->reconfig_posted = 0;

        spin_unlock_bh(&nn->reconfig_lock);

        if (cancelled_timer) {
                del_timer_sync(&nn->reconfig_timer);
                nfp_net_reconfig_wait(nn, nn->reconfig_timer.expires);
        }

        /* Run the posted reconfigs which were issued before we started */
        if (pre_posted_requests) {
                nfp_net_reconfig_start(nn, pre_posted_requests);
                nfp_net_reconfig_wait(nn, jiffies + HZ * NFP_NET_POLL_TIMEOUT);
        }
}

static void nfp_net_reconfig_wait_posted(struct nfp_net *nn)
{
        nfp_net_reconfig_sync_enter(nn);

        spin_lock_bh(&nn->reconfig_lock);
        nn->reconfig_sync_present = false;
        spin_unlock_bh(&nn->reconfig_lock);
}

/**
 * __nfp_net_reconfig() - Reconfigure the firmware
 * @nn:      NFP Net device to reconfigure
 * @update:  The value for the update field in the BAR config
 *
 * Write the update word to the BAR and ping the reconfig queue.  The
 * poll until the firmware has acknowledged the update by zeroing the
 * update word.
 *
 * Return: Negative errno on error, 0 on success
 */
int __nfp_net_reconfig(struct nfp_net *nn, u32 update)
{
        int ret;

        nfp_net_reconfig_sync_enter(nn);

        nfp_net_reconfig_start(nn, update);
        ret = nfp_net_reconfig_wait(nn, jiffies + HZ * NFP_NET_POLL_TIMEOUT);

        spin_lock_bh(&nn->reconfig_lock);

        if (nn->reconfig_posted)
                nfp_net_reconfig_start_async(nn, 0);

        nn->reconfig_sync_present = false;

        spin_unlock_bh(&nn->reconfig_lock);

        return ret;
}

int nfp_net_reconfig(struct nfp_net *nn, u32 update)
{
        int ret;

        nn_ctrl_bar_lock(nn);
        ret = __nfp_net_reconfig(nn, update);
        nn_ctrl_bar_unlock(nn);

        return ret;
}

int nfp_net_mbox_lock(struct nfp_net *nn, unsigned int data_size)
{
        if (nn->tlv_caps.mbox_len < NFP_NET_CFG_MBOX_SIMPLE_VAL + data_size) {
                nn_err(nn, "mailbox too small for %u of data (%u)\n",
                       data_size, nn->tlv_caps.mbox_len);
                return -EIO;
        }

        nn_ctrl_bar_lock(nn);
        return 0;
}

/**
 * nfp_net_mbox_reconfig() - Reconfigure the firmware via the mailbox
 * @nn:        NFP Net device to reconfigure
 * @mbox_cmd:  The value for the mailbox command
 *
 * Helper function for mailbox updates
 *
 * Return: Negative errno on error, 0 on success
 */
int nfp_net_mbox_reconfig(struct nfp_net *nn, u32 mbox_cmd)
{
        u32 mbox = nn->tlv_caps.mbox_off;
        int ret;

        nn_writeq(nn, mbox + NFP_NET_CFG_MBOX_SIMPLE_CMD, mbox_cmd);

        ret = __nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_MBOX);
        if (ret) {
                nn_err(nn, "Mailbox update error\n");
                return ret;
        }

        return -nn_readl(nn, mbox + NFP_NET_CFG_MBOX_SIMPLE_RET);
}

void nfp_net_mbox_reconfig_post(struct nfp_net *nn, u32 mbox_cmd)
{
        u32 mbox = nn->tlv_caps.mbox_off;

        nn_writeq(nn, mbox + NFP_NET_CFG_MBOX_SIMPLE_CMD, mbox_cmd);

        nfp_net_reconfig_post(nn, NFP_NET_CFG_UPDATE_MBOX);
}

int nfp_net_mbox_reconfig_wait_posted(struct nfp_net *nn)
{
        u32 mbox = nn->tlv_caps.mbox_off;

        nfp_net_reconfig_wait_posted(nn);

        return -nn_readl(nn, mbox + NFP_NET_CFG_MBOX_SIMPLE_RET);
}

int nfp_net_mbox_reconfig_and_unlock(struct nfp_net *nn, u32 mbox_cmd)
{
        int ret;

        ret = nfp_net_mbox_reconfig(nn, mbox_cmd);
        nn_ctrl_bar_unlock(nn);
        return ret;
}

/* Interrupt configuration and handling
 */

/**
 * nfp_net_irq_unmask() - Unmask automasked interrupt
 * @nn:       NFP Network structure
 * @entry_nr: MSI-X table entry
 *
 * Clear the ICR for the IRQ entry.
 */
static void nfp_net_irq_unmask(struct nfp_net *nn, unsigned int entry_nr)
{
        nn_writeb(nn, NFP_NET_CFG_ICR(entry_nr), NFP_NET_CFG_ICR_UNMASKED);
        nn_pci_flush(nn);
}

/**
 * nfp_net_irqs_alloc() - allocates MSI-X irqs
 * @pdev:        PCI device structure
 * @irq_entries: Array to be initialized and used to hold the irq entries
 * @min_irqs:    Minimal acceptable number of interrupts
 * @wanted_irqs: Target number of interrupts to allocate
 *
 * Return: Number of irqs obtained or 0 on error.
 */
unsigned int
nfp_net_irqs_alloc(struct pci_dev *pdev, struct msix_entry *irq_entries,
                   unsigned int min_irqs, unsigned int wanted_irqs)
{
        unsigned int i;
        int got_irqs;

        for (i = 0; i < wanted_irqs; i++)
                irq_entries[i].entry = i;

        got_irqs = pci_enable_msix_range(pdev, irq_entries,
                                         min_irqs, wanted_irqs);
        if (got_irqs < 0) {
                dev_err(&pdev->dev, "Failed to enable %d-%d MSI-X (err=%d)\n",
                        min_irqs, wanted_irqs, got_irqs);
                return 0;
        }

        if (got_irqs < wanted_irqs)
                dev_warn(&pdev->dev, "Unable to allocate %d IRQs got only %d\n",
                         wanted_irqs, got_irqs);

        return got_irqs;
}

/**
 * nfp_net_irqs_assign() - Assign interrupts allocated externally to netdev
 * @nn:          NFP Network structure
 * @irq_entries: Table of allocated interrupts
 * @n:           Size of @irq_entries (number of entries to grab)
 *
 * After interrupts are allocated with nfp_net_irqs_alloc() this function
 * should be called to assign them to a specific netdev (port).
 */
void
nfp_net_irqs_assign(struct nfp_net *nn, struct msix_entry *irq_entries,
                    unsigned int n)
{
        struct nfp_net_dp *dp = &nn->dp;

        nn->max_r_vecs = n - NFP_NET_NON_Q_VECTORS;
        dp->num_r_vecs = nn->max_r_vecs;

        memcpy(nn->irq_entries, irq_entries, sizeof(*irq_entries) * n);

        if (dp->num_rx_rings > dp->num_r_vecs ||
            dp->num_tx_rings > dp->num_r_vecs)
                dev_warn(nn->dp.dev, "More rings (%d,%d) than vectors (%d).\n",
                         dp->num_rx_rings, dp->num_tx_rings,
                         dp->num_r_vecs);

        dp->num_rx_rings = min(dp->num_r_vecs, dp->num_rx_rings);
        dp->num_tx_rings = min(dp->num_r_vecs, dp->num_tx_rings);
        dp->num_stack_tx_rings = dp->num_tx_rings;
}

/**
 * nfp_net_irqs_disable() - Disable interrupts
 * @pdev:        PCI device structure
 *
 * Undoes what @nfp_net_irqs_alloc() does.
 */
void nfp_net_irqs_disable(struct pci_dev *pdev)
{
        pci_disable_msix(pdev);
}

/**
 * nfp_net_irq_rxtx() - Interrupt service routine for RX/TX rings.
 * @irq:      Interrupt
 * @data:     Opaque data structure
 *
 * Return: Indicate if the interrupt has been handled.
 */
static irqreturn_t nfp_net_irq_rxtx(int irq, void *data)
{
        struct nfp_net_r_vector *r_vec = data;

        napi_schedule_irqoff(&r_vec->napi);

        /* The FW auto-masks any interrupt, either via the MASK bit in
         * the MSI-X table or via the per entry ICR field.  So there
         * is no need to disable interrupts here.
         */
        return IRQ_HANDLED;
}

static irqreturn_t nfp_ctrl_irq_rxtx(int irq, void *data)
{
        struct nfp_net_r_vector *r_vec = data;

        tasklet_schedule(&r_vec->tasklet);

        return IRQ_HANDLED;
}

/**
 * nfp_net_read_link_status() - Reread link status from control BAR
 * @nn:       NFP Network structure
 */
static void nfp_net_read_link_status(struct nfp_net *nn)
{
        unsigned long flags;
        bool link_up;
        u32 sts;

        spin_lock_irqsave(&nn->link_status_lock, flags);

        sts = nn_readl(nn, NFP_NET_CFG_STS);
        link_up = !!(sts & NFP_NET_CFG_STS_LINK);

        if (nn->link_up == link_up)
                goto out;

        nn->link_up = link_up;
        if (nn->port)
                set_bit(NFP_PORT_CHANGED, &nn->port->flags);

        if (nn->link_up) {
                netif_carrier_on(nn->dp.netdev);
                netdev_info(nn->dp.netdev, "NIC Link is Up\n");
        } else {
                netif_carrier_off(nn->dp.netdev);
                netdev_info(nn->dp.netdev, "NIC Link is Down\n");
        }
out:
        spin_unlock_irqrestore(&nn->link_status_lock, flags);
}

/**
 * nfp_net_irq_lsc() - Interrupt service routine for link state changes
 * @irq:      Interrupt
 * @data:     Opaque data structure
 *
 * Return: Indicate if the interrupt has been handled.
 */
static irqreturn_t nfp_net_irq_lsc(int irq, void *data)
{
        struct nfp_net *nn = data;
        struct msix_entry *entry;

        entry = &nn->irq_entries[NFP_NET_IRQ_LSC_IDX];

        nfp_net_read_link_status(nn);

        nfp_net_irq_unmask(nn, entry->entry);

        return IRQ_HANDLED;
}

/**
 * nfp_net_irq_exn() - Interrupt service routine for exceptions
 * @irq:      Interrupt
 * @data:     Opaque data structure
 *
 * Return: Indicate if the interrupt has been handled.
 */
static irqreturn_t nfp_net_irq_exn(int irq, void *data)
{
        struct nfp_net *nn = data;

        nn_err(nn, "%s: UNIMPLEMENTED.\n", __func__);
        /* XXX TO BE IMPLEMENTED */
        return IRQ_HANDLED;
}

/**
 * nfp_net_tx_ring_init() - Fill in the boilerplate for a TX ring
 * @tx_ring:  TX ring structure
 * @r_vec:    IRQ vector servicing this ring
 * @idx:      Ring index
 * @is_xdp:   Is this an XDP TX ring?
 */
static void
nfp_net_tx_ring_init(struct nfp_net_tx_ring *tx_ring,
                     struct nfp_net_r_vector *r_vec, unsigned int idx,
                     bool is_xdp)
{
        struct nfp_net *nn = r_vec->nfp_net;

        tx_ring->idx = idx;
        tx_ring->r_vec = r_vec;
        tx_ring->is_xdp = is_xdp;
        u64_stats_init(&tx_ring->r_vec->tx_sync);

        tx_ring->qcidx = tx_ring->idx * nn->stride_tx;
        tx_ring->qcp_q = nn->tx_bar + NFP_QCP_QUEUE_OFF(tx_ring->qcidx);
}

/**
 * nfp_net_rx_ring_init() - Fill in the boilerplate for a RX ring
 * @rx_ring:  RX ring structure
 * @r_vec:    IRQ vector servicing this ring
 * @idx:      Ring index
 */
static void
nfp_net_rx_ring_init(struct nfp_net_rx_ring *rx_ring,
                     struct nfp_net_r_vector *r_vec, unsigned int idx)
{
        struct nfp_net *nn = r_vec->nfp_net;

        rx_ring->idx = idx;
        rx_ring->r_vec = r_vec;
        u64_stats_init(&rx_ring->r_vec->rx_sync);

        rx_ring->fl_qcidx = rx_ring->idx * nn->stride_rx;
        rx_ring->qcp_fl = nn->rx_bar + NFP_QCP_QUEUE_OFF(rx_ring->fl_qcidx);
}

/**
 * nfp_net_aux_irq_request() - Request an auxiliary interrupt (LSC or EXN)
 * @nn:         NFP Network structure
 * @ctrl_offset: Control BAR offset where IRQ configuration should be written
 * @format:     printf-style format to construct the interrupt name
 * @name:       Pointer to allocated space for interrupt name
 * @name_sz:    Size of space for interrupt name
 * @vector_idx: Index of MSI-X vector used for this interrupt
 * @handler:    IRQ handler to register for this interrupt
 */
static int
nfp_net_aux_irq_request(struct nfp_net *nn, u32 ctrl_offset,
                        const char *format, char *name, size_t name_sz,
                        unsigned int vector_idx, irq_handler_t handler)
{
        struct msix_entry *entry;
        int err;

        entry = &nn->irq_entries[vector_idx];

        snprintf(name, name_sz, format, nfp_net_name(nn));
        err = request_irq(entry->vector, handler, 0, name, nn);
        if (err) {
                nn_err(nn, "Failed to request IRQ %d (err=%d).\n",
                       entry->vector, err);
                return err;
        }
        nn_writeb(nn, ctrl_offset, entry->entry);
        nfp_net_irq_unmask(nn, entry->entry);

        return 0;
}

/**
 * nfp_net_aux_irq_free() - Free an auxiliary interrupt (LSC or EXN)
 * @nn:         NFP Network structure
 * @ctrl_offset: Control BAR offset where IRQ configuration should be written
 * @vector_idx: Index of MSI-X vector used for this interrupt
 */
static void nfp_net_aux_irq_free(struct nfp_net *nn, u32 ctrl_offset,
                                 unsigned int vector_idx)
{
        nn_writeb(nn, ctrl_offset, 0xff);
        nn_pci_flush(nn);
        free_irq(nn->irq_entries[vector_idx].vector, nn);
}

/* Transmit
 *
 * One queue controller peripheral queue is used for transmit.  The
 * driver en-queues packets for transmit by advancing the write
 * pointer.  The device indicates that packets have transmitted by
 * advancing the read pointer.  The driver maintains a local copy of
 * the read and write pointer in @struct nfp_net_tx_ring.  The driver
 * keeps @wr_p in sync with the queue controller write pointer and can
 * determine how many packets have been transmitted by comparing its
 * copy of the read pointer @rd_p with the read pointer maintained by
 * the queue controller peripheral.
 */

/**
 * nfp_net_tx_full() - Check if the TX ring is full
 * @tx_ring: TX ring to check
 * @dcnt:    Number of descriptors that need to be enqueued (must be >= 1)
 *
 * This function checks, based on the *host copy* of read/write
 * pointer if a given TX ring is full.  The real TX queue may have
 * some newly made available slots.
 *
 * Return: True if the ring is full.
 */
static int nfp_net_tx_full(struct nfp_net_tx_ring *tx_ring, int dcnt)
{
        return (tx_ring->wr_p - tx_ring->rd_p) >= (tx_ring->cnt - dcnt);
}

/* Wrappers for deciding when to stop and restart TX queues */
static int nfp_net_tx_ring_should_wake(struct nfp_net_tx_ring *tx_ring)
{
        return !nfp_net_tx_full(tx_ring, MAX_SKB_FRAGS * 4);
}

static int nfp_net_tx_ring_should_stop(struct nfp_net_tx_ring *tx_ring)
{
        return nfp_net_tx_full(tx_ring, MAX_SKB_FRAGS + 1);
}

/**
 * nfp_net_tx_ring_stop() - stop tx ring
 * @nd_q:    netdev queue
 * @tx_ring: driver tx queue structure
 *
 * Safely stop TX ring.  Remember that while we are running .start_xmit()
 * someone else may be cleaning the TX ring completions so we need to be
 * extra careful here.
 */
static void nfp_net_tx_ring_stop(struct netdev_queue *nd_q,
                                 struct nfp_net_tx_ring *tx_ring)
{
        netif_tx_stop_queue(nd_q);

        /* We can race with the TX completion out of NAPI so recheck */
        smp_mb();
        if (unlikely(nfp_net_tx_ring_should_wake(tx_ring)))
                netif_tx_start_queue(nd_q);
}

/**
 * nfp_net_tx_tso() - Set up Tx descriptor for LSO
 * @r_vec: per-ring structure
 * @txbuf: Pointer to driver soft TX descriptor
 * @txd: Pointer to HW TX descriptor
 * @skb: Pointer to SKB
 * @md_bytes: Prepend length
 *
 * Set up Tx descriptor for LSO, do nothing for non-LSO skbs.
 * Return error on packet header greater than maximum supported LSO header size.
 */
static void nfp_net_tx_tso(struct nfp_net_r_vector *r_vec,
                           struct nfp_net_tx_buf *txbuf,
                           struct nfp_net_tx_desc *txd, struct sk_buff *skb,
                           u32 md_bytes)
{
        u32 l3_offset, l4_offset, hdrlen;
        u16 mss;

        if (!skb_is_gso(skb))
                return;

        if (!skb->encapsulation) {
                l3_offset = skb_network_offset(skb);
                l4_offset = skb_transport_offset(skb);
                hdrlen = skb_transport_offset(skb) + tcp_hdrlen(skb);
        } else {
                l3_offset = skb_inner_network_offset(skb);
                l4_offset = skb_inner_transport_offset(skb);
                hdrlen = skb_inner_transport_header(skb) - skb->data +
                        inner_tcp_hdrlen(skb);
        }

        txbuf->pkt_cnt = skb_shinfo(skb)->gso_segs;
        txbuf->real_len += hdrlen * (txbuf->pkt_cnt - 1);

        mss = skb_shinfo(skb)->gso_size & PCIE_DESC_TX_MSS_MASK;
        txd->l3_offset = l3_offset - md_bytes;
        txd->l4_offset = l4_offset - md_bytes;
        txd->lso_hdrlen = hdrlen - md_bytes;
        txd->mss = cpu_to_le16(mss);
        txd->flags |= PCIE_DESC_TX_LSO;

        u64_stats_update_begin(&r_vec->tx_sync);
        r_vec->tx_lso++;
        u64_stats_update_end(&r_vec->tx_sync);
}

/**
 * nfp_net_tx_csum() - Set TX CSUM offload flags in TX descriptor
 * @dp:  NFP Net data path struct
 * @r_vec: per-ring structure
 * @txbuf: Pointer to driver soft TX descriptor
 * @txd: Pointer to TX descriptor
 * @skb: Pointer to SKB
 *
 * This function sets the TX checksum flags in the TX descriptor based
 * on the configuration and the protocol of the packet to be transmitted.
 */
static void nfp_net_tx_csum(struct nfp_net_dp *dp,
                            struct nfp_net_r_vector *r_vec,
                            struct nfp_net_tx_buf *txbuf,
                            struct nfp_net_tx_desc *txd, struct sk_buff *skb)
{
        struct ipv6hdr *ipv6h;
        struct iphdr *iph;
        u8 l4_hdr;

        if (!(dp->ctrl & NFP_NET_CFG_CTRL_TXCSUM))
                return;

        if (skb->ip_summed != CHECKSUM_PARTIAL)
                return;

        txd->flags |= PCIE_DESC_TX_CSUM;
        if (skb->encapsulation)
                txd->flags |= PCIE_DESC_TX_ENCAP;

        iph = skb->encapsulation ? inner_ip_hdr(skb) : ip_hdr(skb);
        ipv6h = skb->encapsulation ? inner_ipv6_hdr(skb) : ipv6_hdr(skb);

        if (iph->version == 4) {
                txd->flags |= PCIE_DESC_TX_IP4_CSUM;
                l4_hdr = iph->protocol;
        } else if (ipv6h->version == 6) {
                l4_hdr = ipv6h->nexthdr;
        } else {
                nn_dp_warn(dp, "partial checksum but ipv=%x!\n", iph->version);
                return;
        }

        switch (l4_hdr) {
        case IPPROTO_TCP:
                txd->flags |= PCIE_DESC_TX_TCP_CSUM;
                break;
        case IPPROTO_UDP:
                txd->flags |= PCIE_DESC_TX_UDP_CSUM;
                break;
        default:
                nn_dp_warn(dp, "partial checksum but l4 proto=%x!\n", l4_hdr);
                return;
        }

        u64_stats_update_begin(&r_vec->tx_sync);
        if (skb->encapsulation)
                r_vec->hw_csum_tx_inner += txbuf->pkt_cnt;
        else
                r_vec->hw_csum_tx += txbuf->pkt_cnt;
        u64_stats_update_end(&r_vec->tx_sync);
}

static struct sk_buff *
nfp_net_tls_tx(struct nfp_net_dp *dp, struct nfp_net_r_vector *r_vec,
               struct sk_buff *skb, u64 *tls_handle, int *nr_frags)
{
#ifdef CONFIG_TLS_DEVICE
        struct nfp_net_tls_offload_ctx *ntls;
        struct sk_buff *nskb;
        bool resync_pending;
        u32 datalen, seq;

        if (likely(!dp->ktls_tx))
                return skb;
        if (!skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk))
                return skb;

        datalen = skb->len - (skb_transport_offset(skb) + tcp_hdrlen(skb));
        seq = ntohl(tcp_hdr(skb)->seq);
        ntls = tls_driver_ctx(skb->sk, TLS_OFFLOAD_CTX_DIR_TX);
        resync_pending = tls_offload_tx_resync_pending(skb->sk);
        if (unlikely(resync_pending || ntls->next_seq != seq)) {
                /* Pure ACK out of order already */
                if (!datalen)
                        return skb;

                u64_stats_update_begin(&r_vec->tx_sync);
                r_vec->tls_tx_fallback++;
                u64_stats_update_end(&r_vec->tx_sync);

                nskb = tls_encrypt_skb(skb);
                if (!nskb) {
                        u64_stats_update_begin(&r_vec->tx_sync);
                        r_vec->tls_tx_no_fallback++;
                        u64_stats_update_end(&r_vec->tx_sync);
                        return NULL;
                }
                /* encryption wasn't necessary */
                if (nskb == skb)
                        return skb;
                /* we don't re-check ring space */
                if (unlikely(skb_is_nonlinear(nskb))) {
                        nn_dp_warn(dp, "tls_encrypt_skb() produced fragmented frame\n");
                        u64_stats_update_begin(&r_vec->tx_sync);
                        r_vec->tx_errors++;
                        u64_stats_update_end(&r_vec->tx_sync);
                        dev_kfree_skb_any(nskb);
                        return NULL;
                }

                /* jump forward, a TX may have gotten lost, need to sync TX */
                if (!resync_pending && seq - ntls->next_seq < U32_MAX / 4)
                        tls_offload_tx_resync_request(nskb->sk, seq,
                                                      ntls->next_seq);

                *nr_frags = 0;
                return nskb;
        }

        if (datalen) {
                u64_stats_update_begin(&r_vec->tx_sync);
                if (!skb_is_gso(skb))
                        r_vec->hw_tls_tx++;
                else
                        r_vec->hw_tls_tx += skb_shinfo(skb)->gso_segs;
                u64_stats_update_end(&r_vec->tx_sync);
        }

        memcpy(tls_handle, ntls->fw_handle, sizeof(ntls->fw_handle));
        ntls->next_seq += datalen;
#endif
        return skb;
}

static void nfp_net_tls_tx_undo(struct sk_buff *skb, u64 tls_handle)
{
#ifdef CONFIG_TLS_DEVICE
        struct nfp_net_tls_offload_ctx *ntls;
        u32 datalen, seq;

        if (!tls_handle)
                return;
        if (WARN_ON_ONCE(!skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk)))
                return;

        datalen = skb->len - (skb_transport_offset(skb) + tcp_hdrlen(skb));
        seq = ntohl(tcp_hdr(skb)->seq);

        ntls = tls_driver_ctx(skb->sk, TLS_OFFLOAD_CTX_DIR_TX);
        if (ntls->next_seq == seq + datalen)
                ntls->next_seq = seq;
        else
                WARN_ON_ONCE(1);
#endif
}

static void nfp_net_tx_xmit_more_flush(struct nfp_net_tx_ring *tx_ring)
{
        wmb();
        nfp_qcp_wr_ptr_add(tx_ring->qcp_q, tx_ring->wr_ptr_add);
        tx_ring->wr_ptr_add = 0;
}

static int nfp_net_prep_tx_meta(struct sk_buff *skb, u64 tls_handle)
{
        struct metadata_dst *md_dst = skb_metadata_dst(skb);
        unsigned char *data;
        u32 meta_id = 0;
        int md_bytes;

        if (likely(!md_dst && !tls_handle))
                return 0;
        if (unlikely(md_dst && md_dst->type != METADATA_HW_PORT_MUX)) {
                if (!tls_handle)
                        return 0;
                md_dst = NULL;
        }

        md_bytes = 4 + !!md_dst * 4 + !!tls_handle * 8;

        if (unlikely(skb_cow_head(skb, md_bytes)))
                return -ENOMEM;

        meta_id = 0;
        data = skb_push(skb, md_bytes) + md_bytes;
        if (md_dst) {
                data -= 4;
                put_unaligned_be32(md_dst->u.port_info.port_id, data);
                meta_id = NFP_NET_META_PORTID;
        }
        if (tls_handle) {
                /* conn handle is opaque, we just use u64 to be able to quickly
                 * compare it to zero
                 */
                data -= 8;
                memcpy(data, &tls_handle, sizeof(tls_handle));
                meta_id <<= NFP_NET_META_FIELD_SIZE;
                meta_id |= NFP_NET_META_CONN_HANDLE;
        }

        data -= 4;
        put_unaligned_be32(meta_id, data);

        return md_bytes;
}

/**
 * nfp_net_tx() - Main transmit entry point
 * @skb:    SKB to transmit
 * @netdev: netdev structure
 *
 * Return: NETDEV_TX_OK on success.
 */
static netdev_tx_t nfp_net_tx(struct sk_buff *skb, struct net_device *netdev)
{
        struct nfp_net *nn = netdev_priv(netdev);
        const skb_frag_t *frag;
        int f, nr_frags, wr_idx, md_bytes;
        struct nfp_net_tx_ring *tx_ring;
        struct nfp_net_r_vector *r_vec;
        struct nfp_net_tx_buf *txbuf;
        struct nfp_net_tx_desc *txd;
        struct netdev_queue *nd_q;
        struct nfp_net_dp *dp;
        dma_addr_t dma_addr;
        unsigned int fsize;
        u64 tls_handle = 0;
        u16 qidx;

        dp = &nn->dp;
        qidx = skb_get_queue_mapping(skb);
        tx_ring = &dp->tx_rings[qidx];
        r_vec = tx_ring->r_vec;

        nr_frags = skb_shinfo(skb)->nr_frags;

        if (unlikely(nfp_net_tx_full(tx_ring, nr_frags + 1))) {

                nn_dp_warn(dp, "TX ring %d busy. wrp=%u rdp=%u\n",
                           qidx, tx_ring->wr_p, tx_ring->rd_p);
                nd_q = netdev_get_tx_queue(dp->netdev, qidx);
                netif_tx_stop_queue(nd_q);
                nfp_net_tx_xmit_more_flush(tx_ring);
                u64_stats_update_begin(&r_vec->tx_sync);
                r_vec->tx_busy++;
                u64_stats_update_end(&r_vec->tx_sync);
                return NETDEV_TX_BUSY;
        }

        skb = nfp_net_tls_tx(dp, r_vec, skb, &tls_handle, &nr_frags);
        if (unlikely(!skb)) {
                nfp_net_tx_xmit_more_flush(tx_ring);
                return NETDEV_TX_OK;
        }

        md_bytes = nfp_net_prep_tx_meta(skb, tls_handle);
        if (unlikely(md_bytes < 0))
                goto err_flush;

        /* Start with the head skbuf */
        dma_addr = dma_map_single(dp->dev, skb->data, skb_headlen(skb),
                                  DMA_TO_DEVICE);
        if (dma_mapping_error(dp->dev, dma_addr))
                goto err_dma_err;

        wr_idx = D_IDX(tx_ring, tx_ring->wr_p);

        /* Stash the soft descriptor of the head then initialize it */
        txbuf = &tx_ring->txbufs[wr_idx];
        txbuf->skb = skb;
        txbuf->dma_addr = dma_addr;
        txbuf->fidx = -1;
        txbuf->pkt_cnt = 1;
        txbuf->real_len = skb->len;

        /* Build TX descriptor */
        txd = &tx_ring->txds[wr_idx];
        txd->offset_eop = (nr_frags ? 0 : PCIE_DESC_TX_EOP) | md_bytes;
        txd->dma_len = cpu_to_le16(skb_headlen(skb));
        nfp_desc_set_dma_addr(txd, dma_addr);
        txd->data_len = cpu_to_le16(skb->len);

        txd->flags = 0;
        txd->mss = 0;
        txd->lso_hdrlen = 0;

        /* Do not reorder - tso may adjust pkt cnt, vlan may override fields */
        nfp_net_tx_tso(r_vec, txbuf, txd, skb, md_bytes);
        nfp_net_tx_csum(dp, r_vec, txbuf, txd, skb);
        if (skb_vlan_tag_present(skb) && dp->ctrl & NFP_NET_CFG_CTRL_TXVLAN) {
                txd->flags |= PCIE_DESC_TX_VLAN;
                txd->vlan = cpu_to_le16(skb_vlan_tag_get(skb));
        }

        /* Gather DMA */
        if (nr_frags > 0) {
                __le64 second_half;

                /* all descs must match except for in addr, length and eop */
                second_half = txd->vals8[1];

                for (f = 0; f < nr_frags; f++) {
                        frag = &skb_shinfo(skb)->frags[f];
                        fsize = skb_frag_size(frag);

                        dma_addr = skb_frag_dma_map(dp->dev, frag, 0,
                                                    fsize, DMA_TO_DEVICE);
                        if (dma_mapping_error(dp->dev, dma_addr))
                                goto err_unmap;

                        wr_idx = D_IDX(tx_ring, wr_idx + 1);
                        tx_ring->txbufs[wr_idx].skb = skb;
                        tx_ring->txbufs[wr_idx].dma_addr = dma_addr;
                        tx_ring->txbufs[wr_idx].fidx = f;

                        txd = &tx_ring->txds[wr_idx];
                        txd->dma_len = cpu_to_le16(fsize);
                        nfp_desc_set_dma_addr(txd, dma_addr);
                        txd->offset_eop = md_bytes |
                                ((f == nr_frags - 1) ? PCIE_DESC_TX_EOP : 0);
                        txd->vals8[1] = second_half;
                }

                u64_stats_update_begin(&r_vec->tx_sync);
                r_vec->tx_gather++;
                u64_stats_update_end(&r_vec->tx_sync);
        }

        skb_tx_timestamp(skb);

        nd_q = netdev_get_tx_queue(dp->netdev, tx_ring->idx);

        tx_ring->wr_p += nr_frags + 1;
        if (nfp_net_tx_ring_should_stop(tx_ring))
                nfp_net_tx_ring_stop(nd_q, tx_ring);

        tx_ring->wr_ptr_add += nr_frags + 1;
        if (__netdev_tx_sent_queue(nd_q, txbuf->real_len, netdev_xmit_more()))
                nfp_net_tx_xmit_more_flush(tx_ring);

        return NETDEV_TX_OK;

err_unmap:
        while (--f >= 0) {
                frag = &skb_shinfo(skb)->frags[f];
                dma_unmap_page(dp->dev, tx_ring->txbufs[wr_idx].dma_addr,
                               skb_frag_size(frag), DMA_TO_DEVICE);
                tx_ring->txbufs[wr_idx].skb = NULL;
                tx_ring->txbufs[wr_idx].dma_addr = 0;
                tx_ring->txbufs[wr_idx].fidx = -2;
                wr_idx = wr_idx - 1;
                if (wr_idx < 0)
                        wr_idx += tx_ring->cnt;
        }
        dma_unmap_single(dp->dev, tx_ring->txbufs[wr_idx].dma_addr,
                         skb_headlen(skb), DMA_TO_DEVICE);
        tx_ring->txbufs[wr_idx].skb = NULL;
        tx_ring->txbufs[wr_idx].dma_addr = 0;
        tx_ring->txbufs[wr_idx].fidx = -2;
err_dma_err:
        nn_dp_warn(dp, "Failed to map DMA TX buffer\n");
err_flush:
        nfp_net_tx_xmit_more_flush(tx_ring);
        u64_stats_update_begin(&r_vec->tx_sync);
        r_vec->tx_errors++;
        u64_stats_update_end(&r_vec->tx_sync);
        nfp_net_tls_tx_undo(skb, tls_handle);
        dev_kfree_skb_any(skb);
        return NETDEV_TX_OK;
}

/**
 * nfp_net_tx_complete() - Handled completed TX packets
 * @tx_ring:    TX ring structure
 * @budget:     NAPI budget (only used as bool to determine if in NAPI context)
 */
static void nfp_net_tx_complete(struct nfp_net_tx_ring *tx_ring, int budget)
{
        struct nfp_net_r_vector *r_vec = tx_ring->r_vec;
        struct nfp_net_dp *dp = &r_vec->nfp_net->dp;
        struct netdev_queue *nd_q;
        u32 done_pkts = 0, done_bytes = 0;
        u32 qcp_rd_p;
        int todo;

        if (tx_ring->wr_p == tx_ring->rd_p)
                return;

        /* Work out how many descriptors have been transmitted */
        qcp_rd_p = nfp_qcp_rd_ptr_read(tx_ring->qcp_q);

        if (qcp_rd_p == tx_ring->qcp_rd_p)
                return;

        todo = D_IDX(tx_ring, qcp_rd_p - tx_ring->qcp_rd_p);

        while (todo--) {
                const skb_frag_t *frag;
                struct nfp_net_tx_buf *tx_buf;
                struct sk_buff *skb;
                int fidx, nr_frags;
                int idx;

                idx = D_IDX(tx_ring, tx_ring->rd_p++);
                tx_buf = &tx_ring->txbufs[idx];

                skb = tx_buf->skb;
                if (!skb)
                        continue;

                nr_frags = skb_shinfo(skb)->nr_frags;
                fidx = tx_buf->fidx;

                if (fidx == -1) {
                        /* unmap head */
                        dma_unmap_single(dp->dev, tx_buf->dma_addr,
                                         skb_headlen(skb), DMA_TO_DEVICE);

                        done_pkts += tx_buf->pkt_cnt;
                        done_bytes += tx_buf->real_len;
                } else {
                        /* unmap fragment */
                        frag = &skb_shinfo(skb)->frags[fidx];
                        dma_unmap_page(dp->dev, tx_buf->dma_addr,
                                       skb_frag_size(frag), DMA_TO_DEVICE);
                }

                /* check for last gather fragment */
                if (fidx == nr_frags - 1)
                        napi_consume_skb(skb, budget);

                tx_buf->dma_addr = 0;
                tx_buf->skb = NULL;
                tx_buf->fidx = -2;
        }

        tx_ring->qcp_rd_p = qcp_rd_p;

        u64_stats_update_begin(&r_vec->tx_sync);
        r_vec->tx_bytes += done_bytes;
        r_vec->tx_pkts += done_pkts;
        u64_stats_update_end(&r_vec->tx_sync);

        if (!dp->netdev)
                return;

        nd_q = netdev_get_tx_queue(dp->netdev, tx_ring->idx);
        netdev_tx_completed_queue(nd_q, done_pkts, done_bytes);
        if (nfp_net_tx_ring_should_wake(tx_ring)) {
                /* Make sure TX thread will see updated tx_ring->rd_p */
                smp_mb();

                if (unlikely(netif_tx_queue_stopped(nd_q)))
                        netif_tx_wake_queue(nd_q);
        }

        WARN_ONCE(tx_ring->wr_p - tx_ring->rd_p > tx_ring->cnt,
                  "TX ring corruption rd_p=%u wr_p=%u cnt=%u\n",
                  tx_ring->rd_p, tx_ring->wr_p, tx_ring->cnt);
}

static bool nfp_net_xdp_complete(struct nfp_net_tx_ring *tx_ring)
{
        struct nfp_net_r_vector *r_vec = tx_ring->r_vec;
        u32 done_pkts = 0, done_bytes = 0;
        bool done_all;
        int idx, todo;
        u32 qcp_rd_p;

        /* Work out how many descriptors have been transmitted */
        qcp_rd_p = nfp_qcp_rd_ptr_read(tx_ring->qcp_q);

        if (qcp_rd_p == tx_ring->qcp_rd_p)
                return true;

        todo = D_IDX(tx_ring, qcp_rd_p - tx_ring->qcp_rd_p);

        done_all = todo <= NFP_NET_XDP_MAX_COMPLETE;
        todo = min(todo, NFP_NET_XDP_MAX_COMPLETE);

        tx_ring->qcp_rd_p = D_IDX(tx_ring, tx_ring->qcp_rd_p + todo);

        done_pkts = todo;
        while (todo--) {
                idx = D_IDX(tx_ring, tx_ring->rd_p);
                tx_ring->rd_p++;

                done_bytes += tx_ring->txbufs[idx].real_len;
        }

        u64_stats_update_begin(&r_vec->tx_sync);
        r_vec->tx_bytes += done_bytes;
        r_vec->tx_pkts += done_pkts;
        u64_stats_update_end(&r_vec->tx_sync);

        WARN_ONCE(tx_ring->wr_p - tx_ring->rd_p > tx_ring->cnt,
                  "XDP TX ring corruption rd_p=%u wr_p=%u cnt=%u\n",
                  tx_ring->rd_p, tx_ring->wr_p, tx_ring->cnt);

        return done_all;
}

/**
 * nfp_net_tx_ring_reset() - Free any untransmitted buffers and reset pointers
 * @dp:         NFP Net data path struct
 * @tx_ring:    TX ring structure
 *
 * Assumes that the device is stopped, must be idempotent.
 */
static void
nfp_net_tx_ring_reset(struct nfp_net_dp *dp, struct nfp_net_tx_ring *tx_ring)
{
        const skb_frag_t *frag;
        struct netdev_queue *nd_q;

        while (!tx_ring->is_xdp && tx_ring->rd_p != tx_ring->wr_p) {
                struct nfp_net_tx_buf *tx_buf;
                struct sk_buff *skb;
                int idx, nr_frags;

                idx = D_IDX(tx_ring, tx_ring->rd_p);
                tx_buf = &tx_ring->txbufs[idx];

                skb = tx_ring->txbufs[idx].skb;
                nr_frags = skb_shinfo(skb)->nr_frags;

                if (tx_buf->fidx == -1) {
                        /* unmap head */
                        dma_unmap_single(dp->dev, tx_buf->dma_addr,
                                         skb_headlen(skb), DMA_TO_DEVICE);
                } else {
                        /* unmap fragment */
                        frag = &skb_shinfo(skb)->frags[tx_buf->fidx];
                        dma_unmap_page(dp->dev, tx_buf->dma_addr,
                                       skb_frag_size(frag), DMA_TO_DEVICE);
                }

                /* check for last gather fragment */
                if (tx_buf->fidx == nr_frags - 1)
                        dev_kfree_skb_any(skb);

                tx_buf->dma_addr = 0;
                tx_buf->skb = NULL;
                tx_buf->fidx = -2;

                tx_ring->qcp_rd_p++;
                tx_ring->rd_p++;
        }

        memset(tx_ring->txds, 0, tx_ring->size);
        tx_ring->wr_p = 0;
        tx_ring->rd_p = 0;
        tx_ring->qcp_rd_p = 0;
        tx_ring->wr_ptr_add = 0;

        if (tx_ring->is_xdp || !dp->netdev)
                return;

        nd_q = netdev_get_tx_queue(dp->netdev, tx_ring->idx);
        netdev_tx_reset_queue(nd_q);
}

static void nfp_net_tx_timeout(struct net_device *netdev, unsigned int txqueue)
{
        struct nfp_net *nn = netdev_priv(netdev);

        nn_warn(nn, "TX watchdog timeout on ring: %u\n", txqueue);
}

/* Receive processing
 */
static unsigned int
nfp_net_calc_fl_bufsz(struct nfp_net_dp *dp)
{
        unsigned int fl_bufsz;

        fl_bufsz = NFP_NET_RX_BUF_HEADROOM;
        fl_bufsz += dp->rx_dma_off;
        if (dp->rx_offset == NFP_NET_CFG_RX_OFFSET_DYNAMIC)
                fl_bufsz += NFP_NET_MAX_PREPEND;
        else
                fl_bufsz += dp->rx_offset;
        fl_bufsz += ETH_HLEN + VLAN_HLEN * 2 + dp->mtu;

        fl_bufsz = SKB_DATA_ALIGN(fl_bufsz);
        fl_bufsz += SKB_DATA_ALIGN(sizeof(struct skb_shared_info));

        return fl_bufsz;
}

static void
nfp_net_free_frag(void *frag, bool xdp)
{
        if (!xdp)
                skb_free_frag(frag);
        else
                __free_page(virt_to_page(frag));
}

/**
 * nfp_net_rx_alloc_one() - Allocate and map page frag for RX
 * @dp:         NFP Net data path struct
 * @dma_addr:   Pointer to storage for DMA address (output param)
 *
 * This function will allcate a new page frag, map it for DMA.
 *
 * Return: allocated page frag or NULL on failure.
 */
static void *nfp_net_rx_alloc_one(struct nfp_net_dp *dp, dma_addr_t *dma_addr)
{
        void *frag;

        if (!dp->xdp_prog) {
                frag = netdev_alloc_frag(dp->fl_bufsz);
        } else {
                struct page *page;

                page = alloc_page(GFP_KERNEL);
                frag = page ? page_address(page) : NULL;
        }
        if (!frag) {
                nn_dp_warn(dp, "Failed to alloc receive page frag\n");
                return NULL;
        }

        *dma_addr = nfp_net_dma_map_rx(dp, frag);
        if (dma_mapping_error(dp->dev, *dma_addr)) {
                nfp_net_free_frag(frag, dp->xdp_prog);
                nn_dp_warn(dp, "Failed to map DMA RX buffer\n");
                return NULL;
        }

        return frag;
}

static void *nfp_net_napi_alloc_one(struct nfp_net_dp *dp, dma_addr_t *dma_addr)
{
        void *frag;

        if (!dp->xdp_prog) {
                frag = napi_alloc_frag(dp->fl_bufsz);
                if (unlikely(!frag))
                        return NULL;
        } else {
                struct page *page;

                page = dev_alloc_page();
                if (unlikely(!page))
                        return NULL;
                frag = page_address(page);
        }

        *dma_addr = nfp_net_dma_map_rx(dp, frag);
        if (dma_mapping_error(dp->dev, *dma_addr)) {
                nfp_net_free_frag(frag, dp->xdp_prog);
                nn_dp_warn(dp, "Failed to map DMA RX buffer\n");
                return NULL;
        }

        return frag;
}

/**
 * nfp_net_rx_give_one() - Put mapped skb on the software and hardware rings
 * @dp:         NFP Net data path struct
 * @rx_ring:    RX ring structure
 * @frag:       page fragment buffer
 * @dma_addr:   DMA address of skb mapping
 */
static void nfp_net_rx_give_one(const struct nfp_net_dp *dp,
                                struct nfp_net_rx_ring *rx_ring,
                                void *frag, dma_addr_t dma_addr)
{
        unsigned int wr_idx;

        wr_idx = D_IDX(rx_ring, rx_ring->wr_p);

        nfp_net_dma_sync_dev_rx(dp, dma_addr);

        /* Stash SKB and DMA address away */
        rx_ring->rxbufs[wr_idx].frag = frag;
        rx_ring->rxbufs[wr_idx].dma_addr = dma_addr;

        /* Fill freelist descriptor */
        rx_ring->rxds[wr_idx].fld.reserved = 0;
        rx_ring->rxds[wr_idx].fld.meta_len_dd = 0;
        nfp_desc_set_dma_addr(&rx_ring->rxds[wr_idx].fld,
                              dma_addr + dp->rx_dma_off);

        rx_ring->wr_p++;
        if (!(rx_ring->wr_p % NFP_NET_FL_BATCH)) {
                /* Update write pointer of the freelist queue. Make
                 * sure all writes are flushed before telling the hardware.
                 */
                wmb();
                nfp_qcp_wr_ptr_add(rx_ring->qcp_fl, NFP_NET_FL_BATCH);
        }
}

/**
 * nfp_net_rx_ring_reset() - Reflect in SW state of freelist after disable
 * @rx_ring:    RX ring structure
 *
 * Assumes that the device is stopped, must be idempotent.
 */
static void nfp_net_rx_ring_reset(struct nfp_net_rx_ring *rx_ring)
{
        unsigned int wr_idx, last_idx;

        /* wr_p == rd_p means ring was never fed FL bufs.  RX rings are always
         * kept at cnt - 1 FL bufs.
         */
        if (rx_ring->wr_p == 0 && rx_ring->rd_p == 0)
                return;

        /* Move the empty entry to the end of the list */
        wr_idx = D_IDX(rx_ring, rx_ring->wr_p);
        last_idx = rx_ring->cnt - 1;
        rx_ring->rxbufs[wr_idx].dma_addr = rx_ring->rxbufs[last_idx].dma_addr;
        rx_ring->rxbufs[wr_idx].frag = rx_ring->rxbufs[last_idx].frag;
        rx_ring->rxbufs[last_idx].dma_addr = 0;
        rx_ring->rxbufs[last_idx].frag = NULL;

        memset(rx_ring->rxds, 0, rx_ring->size);
        rx_ring->wr_p = 0;
        rx_ring->rd_p = 0;
}

/**
 * nfp_net_rx_ring_bufs_free() - Free any buffers currently on the RX ring
 * @dp:         NFP Net data path struct
 * @rx_ring:    RX ring to remove buffers from
 *
 * Assumes that the device is stopped and buffers are in [0, ring->cnt - 1)
 * entries.  After device is disabled nfp_net_rx_ring_reset() must be called
 * to restore required ring geometry.
 */
static void
nfp_net_rx_ring_bufs_free(struct nfp_net_dp *dp,
                          struct nfp_net_rx_ring *rx_ring)
{
        unsigned int i;

        for (i = 0; i < rx_ring->cnt - 1; i++) {
                /* NULL skb can only happen when initial filling of the ring
                 * fails to allocate enough buffers and calls here to free
                 * already allocated ones.
                 */
                if (!rx_ring->rxbufs[i].frag)
                        continue;

                nfp_net_dma_unmap_rx(dp, rx_ring->rxbufs[i].dma_addr);
                nfp_net_free_frag(rx_ring->rxbufs[i].frag, dp->xdp_prog);
                rx_ring->rxbufs[i].dma_addr = 0;
                rx_ring->rxbufs[i].frag = NULL;
        }
}

/**
 * nfp_net_rx_ring_bufs_alloc() - Fill RX ring with buffers (don't give to FW)
 * @dp:         NFP Net data path struct
 * @rx_ring:    RX ring to remove buffers from
 */
static int
nfp_net_rx_ring_bufs_alloc(struct nfp_net_dp *dp,
                           struct nfp_net_rx_ring *rx_ring)
{
        struct nfp_net_rx_buf *rxbufs;
        unsigned int i;

        rxbufs = rx_ring->rxbufs;

        for (i = 0; i < rx_ring->cnt - 1; i++) {
                rxbufs[i].frag = nfp_net_rx_alloc_one(dp, &rxbufs[i].dma_addr);
                if (!rxbufs[i].frag) {
                        nfp_net_rx_ring_bufs_free(dp, rx_ring);
                        return -ENOMEM;
                }
        }

        return 0;
}

/**
 * nfp_net_rx_ring_fill_freelist() - Give buffers from the ring to FW
 * @dp:      NFP Net data path struct
 * @rx_ring: RX ring to fill
 */
static void
nfp_net_rx_ring_fill_freelist(struct nfp_net_dp *dp,
                              struct nfp_net_rx_ring *rx_ring)
{
        unsigned int i;

        for (i = 0; i < rx_ring->cnt - 1; i++)
                nfp_net_rx_give_one(dp, rx_ring, rx_ring->rxbufs[i].frag,
                                    rx_ring->rxbufs[i].dma_addr);
}

/**
 * nfp_net_rx_csum_has_errors() - group check if rxd has any csum errors
 * @flags: RX descriptor flags field in CPU byte order
 */
static int nfp_net_rx_csum_has_errors(u16 flags)
{
        u16 csum_all_checked, csum_all_ok;

        csum_all_checked = flags & __PCIE_DESC_RX_CSUM_ALL;
        csum_all_ok = flags & __PCIE_DESC_RX_CSUM_ALL_OK;

        return csum_all_checked != (csum_all_ok << PCIE_DESC_RX_CSUM_OK_SHIFT);
}

/**
 * nfp_net_rx_csum() - set SKB checksum field based on RX descriptor flags
 * @dp:  NFP Net data path struct
 * @r_vec: per-ring structure
 * @rxd: Pointer to RX descriptor
 * @meta: Parsed metadata prepend
 * @skb: Pointer to SKB
 */
static void nfp_net_rx_csum(struct nfp_net_dp *dp,
                            struct nfp_net_r_vector *r_vec,
                            struct nfp_net_rx_desc *rxd,
                            struct nfp_meta_parsed *meta, struct sk_buff *skb)
{
        skb_checksum_none_assert(skb);

        if (!(dp->netdev->features & NETIF_F_RXCSUM))
                return;

        if (meta->csum_type) {
                skb->ip_summed = meta->csum_type;
                skb->csum = meta->csum;
                u64_stats_update_begin(&r_vec->rx_sync);
                r_vec->hw_csum_rx_complete++;
                u64_stats_update_end(&r_vec->rx_sync);
                return;
        }

        if (nfp_net_rx_csum_has_errors(le16_to_cpu(rxd->rxd.flags))) {
                u64_stats_update_begin(&r_vec->rx_sync);
                r_vec->hw_csum_rx_error++;
                u64_stats_update_end(&r_vec->rx_sync);
                return;
        }

        /* Assume that the firmware will never report inner CSUM_OK unless outer
         * L4 headers were successfully parsed. FW will always report zero UDP
         * checksum as CSUM_OK.
         */
        if (rxd->rxd.flags & PCIE_DESC_RX_TCP_CSUM_OK ||
            rxd->rxd.flags & PCIE_DESC_RX_UDP_CSUM_OK) {
                __skb_incr_checksum_unnecessary(skb);
                u64_stats_update_begin(&r_vec->rx_sync);
                r_vec->hw_csum_rx_ok++;
                u64_stats_update_end(&r_vec->rx_sync);
        }

        if (rxd->rxd.flags & PCIE_DESC_RX_I_TCP_CSUM_OK ||
            rxd->rxd.flags & PCIE_DESC_RX_I_UDP_CSUM_OK) {
                __skb_incr_checksum_unnecessary(skb);
                u64_stats_update_begin(&r_vec->rx_sync);
                r_vec->hw_csum_rx_inner_ok++;
                u64_stats_update_end(&r_vec->rx_sync);
        }
}

static void
nfp_net_set_hash(struct net_device *netdev, struct nfp_meta_parsed *meta,
                 unsigned int type, __be32 *hash)
{
        if (!(netdev->features & NETIF_F_RXHASH))
                return;

        switch (type) {
        case NFP_NET_RSS_IPV4:
        case NFP_NET_RSS_IPV6:
        case NFP_NET_RSS_IPV6_EX:
                meta->hash_type = PKT_HASH_TYPE_L3;
                break;
        default:
                meta->hash_type = PKT_HASH_TYPE_L4;
                break;
        }

        meta->hash = get_unaligned_be32(hash);
}

static void
nfp_net_set_hash_desc(struct net_device *netdev, struct nfp_meta_parsed *meta,
                      void *data, struct nfp_net_rx_desc *rxd)
{
        struct nfp_net_rx_hash *rx_hash = data;

        if (!(rxd->rxd.flags & PCIE_DESC_RX_RSS))
                return;

        nfp_net_set_hash(netdev, meta, get_unaligned_be32(&rx_hash->hash_type),
                         &rx_hash->hash);
}

static bool
nfp_net_parse_meta(struct net_device *netdev, struct nfp_meta_parsed *meta,
                   void *data, void *pkt, unsigned int pkt_len, int meta_len)
{
        u32 meta_info;

        meta_info = get_unaligned_be32(data);
        data += 4;

        while (meta_info) {
                switch (meta_info & NFP_NET_META_FIELD_MASK) {
                case NFP_NET_META_HASH:
                        meta_info >>= NFP_NET_META_FIELD_SIZE;
                        nfp_net_set_hash(netdev, meta,
                                         meta_info & NFP_NET_META_FIELD_MASK,
                                         (__be32 *)data);
                        data += 4;
                        break;
                case NFP_NET_META_MARK:
                        meta->mark = get_unaligned_be32(data);
                        data += 4;
                        break;
                case NFP_NET_META_PORTID:
                        meta->portid = get_unaligned_be32(data);
                        data += 4;
                        break;
                case NFP_NET_META_CSUM:
                        meta->csum_type = CHECKSUM_COMPLETE;
                        meta->csum =
                                (__force __wsum)__get_unaligned_cpu32(data);
                        data += 4;
                        break;
                case NFP_NET_META_RESYNC_INFO:
                        if (nfp_net_tls_rx_resync_req(netdev, data, pkt,
                                                      pkt_len))
                                return NULL;
                        data += sizeof(struct nfp_net_tls_resync_req);
                        break;
                default:
                        return true;
                }

                meta_info >>= NFP_NET_META_FIELD_SIZE;
        }

        return data != pkt;
}

static void
nfp_net_rx_drop(const struct nfp_net_dp *dp, struct nfp_net_r_vector *r_vec,
                struct nfp_net_rx_ring *rx_ring, struct nfp_net_rx_buf *rxbuf,
                struct sk_buff *skb)
{
        u64_stats_update_begin(&r_vec->rx_sync);
        r_vec->rx_drops++;
        /* If we have both skb and rxbuf the replacement buffer allocation
         * must have failed, count this as an alloc failure.
         */
        if (skb && rxbuf)
                r_vec->rx_replace_buf_alloc_fail++;
        u64_stats_update_end(&r_vec->rx_sync);

        /* skb is build based on the frag, free_skb() would free the frag
         * so to be able to reuse it we need an extra ref.
         */
        if (skb && rxbuf && skb->head == rxbuf->frag)
                page_ref_inc(virt_to_head_page(rxbuf->frag));
        if (rxbuf)
                nfp_net_rx_give_one(dp, rx_ring, rxbuf->frag, rxbuf->dma_addr);
        if (skb)
                dev_kfree_skb_any(skb);
}

static bool
nfp_net_tx_xdp_buf(struct nfp_net_dp *dp, struct nfp_net_rx_ring *rx_ring,
                   struct nfp_net_tx_ring *tx_ring,
                   struct nfp_net_rx_buf *rxbuf, unsigned int dma_off,
                   unsigned int pkt_len, bool *completed)
{
        unsigned int dma_map_sz = dp->fl_bufsz - NFP_NET_RX_BUF_NON_DATA;
        struct nfp_net_tx_buf *txbuf;
        struct nfp_net_tx_desc *txd;
        int wr_idx;

        /* Reject if xdp_adjust_tail grow packet beyond DMA area */
        if (pkt_len + dma_off > dma_map_sz)
                return false;

        if (unlikely(nfp_net_tx_full(tx_ring, 1))) {
                if (!*completed) {
                        nfp_net_xdp_complete(tx_ring);
                        *completed = true;
                }

                if (unlikely(nfp_net_tx_full(tx_ring, 1))) {
                        nfp_net_rx_drop(dp, rx_ring->r_vec, rx_ring, rxbuf,
                                        NULL);
                        return false;
                }
        }

        wr_idx = D_IDX(tx_ring, tx_ring->wr_p);

        /* Stash the soft descriptor of the head then initialize it */
        txbuf = &tx_ring->txbufs[wr_idx];

        nfp_net_rx_give_one(dp, rx_ring, txbuf->frag, txbuf->dma_addr);

        txbuf->frag = rxbuf->frag;
        txbuf->dma_addr = rxbuf->dma_addr;
        txbuf->fidx = -1;
        txbuf->pkt_cnt = 1;
        txbuf->real_len = pkt_len;

        dma_sync_single_for_device(dp->dev, rxbuf->dma_addr + dma_off,
                                   pkt_len, DMA_BIDIRECTIONAL);

        /* Build TX descriptor */
        txd = &tx_ring->txds[wr_idx];
        txd->offset_eop = PCIE_DESC_TX_EOP;
        txd->dma_len = cpu_to_le16(pkt_len);
        nfp_desc_set_dma_addr(txd, rxbuf->dma_addr + dma_off);
        txd->data_len = cpu_to_le16(pkt_len);

        txd->flags = 0;
        txd->mss = 0;
        txd->lso_hdrlen = 0;

        tx_ring->wr_p++;
        tx_ring->wr_ptr_add++;
        return true;
}

/**
 * nfp_net_rx() - receive up to @budget packets on @rx_ring
 * @rx_ring:   RX ring to receive from
 * @budget:    NAPI budget
 *
 * Note, this function is separated out from the napi poll function to
 * more cleanly separate packet receive code from other bookkeeping
 * functions performed in the napi poll function.
 *
 * Return: Number of packets received.
 */
static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
{
        struct nfp_net_r_vector *r_vec = rx_ring->r_vec;
        struct nfp_net_dp *dp = &r_vec->nfp_net->dp;
        struct nfp_net_tx_ring *tx_ring;
        struct bpf_prog *xdp_prog;
        bool xdp_tx_cmpl = false;
        unsigned int true_bufsz;
        struct sk_buff *skb;
        int pkts_polled = 0;
        struct xdp_buff xdp;
        int idx;

        xdp_prog = READ_ONCE(dp->xdp_prog);
        true_bufsz = xdp_prog ? PAGE_SIZE : dp->fl_bufsz;
        xdp_init_buff(&xdp, PAGE_SIZE - NFP_NET_RX_BUF_HEADROOM,
                      &rx_ring->xdp_rxq);
        tx_ring = r_vec->xdp_ring;

        while (pkts_polled < budget) {
                unsigned int meta_len, data_len, meta_off, pkt_len, pkt_off;
                struct nfp_net_rx_buf *rxbuf;
                struct nfp_net_rx_desc *rxd;
                struct nfp_meta_parsed meta;
                bool redir_egress = false;
                struct net_device *netdev;
                dma_addr_t new_dma_addr;
                u32 meta_len_xdp = 0;
                void *new_frag;

                idx = D_IDX(rx_ring, rx_ring->rd_p);

                rxd = &rx_ring->rxds[idx];
                if (!(rxd->rxd.meta_len_dd & PCIE_DESC_RX_DD))
                        break;

                /* Memory barrier to ensure that we won't do other reads
                 * before the DD bit.
                 */
                dma_rmb();

                memset(&meta, 0, sizeof(meta));

                rx_ring->rd_p++;
                pkts_polled++;

                rxbuf = &rx_ring->rxbufs[idx];
                /*         < meta_len >
                 *  <-- [rx_offset] -->
                 *  ---------------------------------------------------------
                 * | [XX] |  metadata  |             packet           | XXXX |
                 *  ---------------------------------------------------------
                 *         <---------------- data_len --------------->
                 *
                 * The rx_offset is fixed for all packets, the meta_len can vary
                 * on a packet by packet basis. If rx_offset is set to zero
                 * (_RX_OFFSET_DYNAMIC) metadata starts at the beginning of the
                 * buffer and is immediately followed by the packet (no [XX]).
                 */
                meta_len = rxd->rxd.meta_len_dd & PCIE_DESC_RX_META_LEN_MASK;
                data_len = le16_to_cpu(rxd->rxd.data_len);
                pkt_len = data_len - meta_len;

                pkt_off = NFP_NET_RX_BUF_HEADROOM + dp->rx_dma_off;
                if (dp->rx_offset == NFP_NET_CFG_RX_OFFSET_DYNAMIC)
                        pkt_off += meta_len;
                else
                        pkt_off += dp->rx_offset;
                meta_off = pkt_off - meta_len;

                /* Stats update */
                u64_stats_update_begin(&r_vec->rx_sync);
                r_vec->rx_pkts++;
                r_vec->rx_bytes += pkt_len;
                u64_stats_update_end(&r_vec->rx_sync);

                if (unlikely(meta_len > NFP_NET_MAX_PREPEND ||
                             (dp->rx_offset && meta_len > dp->rx_offset))) {
                        nn_dp_warn(dp, "oversized RX packet metadata %u\n",
                                   meta_len);
                        nfp_net_rx_drop(dp, r_vec, rx_ring, rxbuf, NULL);
                        continue;
                }

                nfp_net_dma_sync_cpu_rx(dp, rxbuf->dma_addr + meta_off,
                                        data_len);

                if (!dp->chained_metadata_format) {
                        nfp_net_set_hash_desc(dp->netdev, &meta,
                                              rxbuf->frag + meta_off, rxd);
                } else if (meta_len) {
                        if (unlikely(nfp_net_parse_meta(dp->netdev, &meta,
                                                        rxbuf->frag + meta_off,
                                                        rxbuf->frag + pkt_off,
                                                        pkt_len, meta_len))) {
                                nn_dp_warn(dp, "invalid RX packet metadata\n");
                                nfp_net_rx_drop(dp, r_vec, rx_ring, rxbuf,
                                                NULL);
                                continue;
                        }
                }

                if (xdp_prog && !meta.portid) {
                        void *orig_data = rxbuf->frag + pkt_off;
                        unsigned int dma_off;
                        int act;

                        xdp_prepare_buff(&xdp,
                                         rxbuf->frag + NFP_NET_RX_BUF_HEADROOM,
                                         pkt_off - NFP_NET_RX_BUF_HEADROOM,
                                         pkt_len, true);

                        act = bpf_prog_run_xdp(xdp_prog, &xdp);

                        pkt_len = xdp.data_end - xdp.data;
                        pkt_off += xdp.data - orig_data;

                        switch (act) {
                        case XDP_PASS:
                                meta_len_xdp = xdp.data - xdp.data_meta;
                                break;
                        case XDP_TX:
                                dma_off = pkt_off - NFP_NET_RX_BUF_HEADROOM;
                                if (unlikely(!nfp_net_tx_xdp_buf(dp, rx_ring,
                                                                 tx_ring, rxbuf,
                                                                 dma_off,
                                                                 pkt_len,
                                                                 &xdp_tx_cmpl)))
                                        trace_xdp_exception(dp->netdev,
                                                            xdp_prog, act);
                                continue;
                        default:
                                bpf_warn_invalid_xdp_action(act);
                                fallthrough;
                        case XDP_ABORTED:
                                trace_xdp_exception(dp->netdev, xdp_prog, act);
                                fallthrough;
                        case XDP_DROP:
                                nfp_net_rx_give_one(dp, rx_ring, rxbuf->frag,
                                                    rxbuf->dma_addr);
                                continue;
                        }
                }

                if (likely(!meta.portid)) {
                        netdev = dp->netdev;
                } else if (meta.portid == NFP_META_PORT_ID_CTRL) {
                        struct nfp_net *nn = netdev_priv(dp->netdev);

                        nfp_app_ctrl_rx_raw(nn->app, rxbuf->frag + pkt_off,
                                            pkt_len);
                        nfp_net_rx_give_one(dp, rx_ring, rxbuf->frag,
                                            rxbuf->dma_addr);
                        continue;
                } else {
                        struct nfp_net *nn;

                        nn = netdev_priv(dp->netdev);
                        netdev = nfp_app_dev_get(nn->app, meta.portid,
                                                 &redir_egress);
                        if (unlikely(!netdev)) {
                                nfp_net_rx_drop(dp, r_vec, rx_ring, rxbuf,
                                                NULL);
                                continue;
                        }

                        if (nfp_netdev_is_nfp_repr(netdev))
                                nfp_repr_inc_rx_stats(netdev, pkt_len);
                }

                skb = build_skb(rxbuf->frag, true_bufsz);
                if (unlikely(!skb)) {
                        nfp_net_rx_drop(dp, r_vec, rx_ring, rxbuf, NULL);
                        continue;
                }
                new_frag = nfp_net_napi_alloc_one(dp, &new_dma_addr);
                if (unlikely(!new_frag)) {
                        nfp_net_rx_drop(dp, r_vec, rx_ring, rxbuf, skb);
                        continue;
                }

                nfp_net_dma_unmap_rx(dp, rxbuf->dma_addr);

                nfp_net_rx_give_one(dp, rx_ring, new_frag, new_dma_addr);

                skb_reserve(skb, pkt_off);
                skb_put(skb, pkt_len);

                skb->mark = meta.mark;
                skb_set_hash(skb, meta.hash, meta.hash_type);

                skb_record_rx_queue(skb, rx_ring->idx);