int mlx4_en_free_tx_buf(struct ether *dev, struct mlx4_en_tx_ring *ring)
{
panic("Disabled");
#if 0 // AKAROS_PORT
struct mlx4_en_priv *priv = netdev_priv(dev);
int cnt = 0;
/* Skip last polled descriptor */
ring->cons += ring->last_nr_txbb;
en_dbg(DRV, priv, "Freeing Tx buf - cons:0x%x prod:0x%x\n",
ring->cons, ring->prod);
if ((uint32_t) (ring->prod - ring->cons) > ring->size) {
if (netif_msg_tx_err(priv))
en_warn(priv, "Tx consumer passed producer!\n");
return 0;
}
while (ring->cons != ring->prod) {
ring->last_nr_txbb = mlx4_en_free_tx_desc(priv, ring,
ring->cons & ring->size_mask,
!!(ring->cons & ring->size), 0);
ring->cons += ring->last_nr_txbb;
cnt++;
}
netdev_tx_reset_queue(ring->tx_queue);
if (cnt)
en_dbg(DRV, priv, "Freed %d uncompleted tx descriptors\n", cnt);
return cnt;
#endif
}
int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
int cnt = 0;
/* Skip last polled descriptor */
ring->cons += ring->last_nr_txbb;
en_dbg(DRV, priv, "Freeing Tx buf - cons:0x%x prod:0x%x\n",
ring->cons, ring->prod);
if ((u32) (ring->prod - ring->cons) > ring->size) {
en_warn(priv, "Tx consumer passed producer!\n");
return 0;
}
while (ring->cons != ring->prod) {
ring->last_nr_txbb = mlx4_en_free_tx_desc(priv, ring,
ring->cons & ring->size_mask,
!!(ring->cons & ring->size), 0);
ring->cons += ring->last_nr_txbb;
cnt++;
}
if (cnt)
en_dbg(DRV, priv, "Freed %d uncompleted tx descriptors\n", cnt);
return cnt;
}
static int get_real_size(struct sk_buff *skb, struct net_device *dev,
int *lso_header_size)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
int real_size;
if (skb_is_gso(skb)) {
*lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb);
real_size = CTRL_SIZE + skb_shinfo(skb)->nr_frags * DS_SIZE +
ALIGN(*lso_header_size + 4, DS_SIZE);
if (unlikely(*lso_header_size != skb_headlen(skb))) {
/* We add a segment for the skb linear buffer only if
* it contains data */
if (*lso_header_size < skb_headlen(skb))
real_size += DS_SIZE;
else {
if (netif_msg_tx_err(priv))
en_warn(priv, "Non-linear headers\n");
return 0;
}
}
} else {
*lso_header_size = 0;
if (!is_inline(skb, NULL))
real_size = CTRL_SIZE + (skb_shinfo(skb)->nr_frags + 1) * DS_SIZE;
else
real_size = inline_size(skb);
}
return real_size;
}
static int mlx4_en_fill_rx_buffers(struct mlx4_en_priv *priv)
{
struct mlx4_en_rx_ring *ring;
int ring_ind;
int buf_ind;
int new_size;
int err;
for (buf_ind = 0; buf_ind < priv->prof->rx_ring_size; buf_ind++) {
for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
ring = priv->rx_ring[ring_ind];
err = mlx4_en_prepare_rx_desc(priv, ring,
ring->actual_size);
if (err) {
if (ring->actual_size == 0) {
en_err(priv, "Failed to allocate "
"enough rx buffers\n");
return -ENOMEM;
} else {
new_size =
rounddown_pow_of_two(ring->actual_size);
en_warn(priv, "Only %d buffers allocated "
"reducing ring size to %d\n",
ring->actual_size, new_size);
goto reduce_rings;
}
}
ring->actual_size++;
ring->prod++;
}
}
return 0;
reduce_rings:
for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
ring = priv->rx_ring[ring_ind];
while (ring->actual_size > new_size) {
ring->actual_size--;
ring->prod--;
mlx4_en_free_buf(ring,
ring->mbuf + ring->actual_size);
}
}
return 0;
}
static int get_real_size(const struct sk_buff *skb,
const struct skb_shared_info *shinfo,
struct net_device *dev,
int *lso_header_size,
bool *inline_ok,
void **pfrag)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
int real_size;
if (shinfo->gso_size) {
*inline_ok = false;
if (skb->encapsulation)
*lso_header_size = (skb_inner_transport_header(skb) - skb->data) + inner_tcp_hdrlen(skb);
else
*lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb);
real_size = CTRL_SIZE + shinfo->nr_frags * DS_SIZE +
ALIGN(*lso_header_size + 4, DS_SIZE);
if (unlikely(*lso_header_size != skb_headlen(skb))) {
/* We add a segment for the skb linear buffer only if
* it contains data */
if (*lso_header_size < skb_headlen(skb))
real_size += DS_SIZE;
else {
if (netif_msg_tx_err(priv))
en_warn(priv, "Non-linear headers\n");
return 0;
}
}
} else {
*lso_header_size = 0;
*inline_ok = is_inline(priv->prof->inline_thold, skb,
shinfo, pfrag);
if (*inline_ok)
real_size = inline_size(skb);
else
real_size = CTRL_SIZE +
(shinfo->nr_frags + 1) * DS_SIZE;
}
return real_size;
}
static int mlx4_en_fill_rx_buffers(struct mlx4_en_priv *priv)
{
struct mlx4_en_rx_ring *ring;
int ring_ind;
int buf_ind;
int new_size;
for (buf_ind = 0; buf_ind < priv->prof->rx_ring_size; buf_ind++) {
for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
ring = priv->rx_ring[ring_ind];
if (mlx4_en_prepare_rx_desc(priv, ring,
ring->actual_size,
GFP_KERNEL | __GFP_COLD)) {
if (ring->actual_size < MLX4_EN_MIN_RX_SIZE) {
en_err(priv, "Failed to allocate enough rx buffers\n");
return -ENOMEM;
} else {
new_size = rounddown_pow_of_two(ring->actual_size);
en_warn(priv, "Only %d buffers allocated reducing ring size to %d\n",
ring->actual_size, new_size);
goto reduce_rings;
}
}
ring->actual_size++;
ring->prod++;
}
}
return 0;
reduce_rings:
for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
ring = priv->rx_ring[ring_ind];
while (ring->actual_size > new_size) {
ring->actual_size--;
ring->prod--;
mlx4_en_free_rx_desc(priv, ring, ring->actual_size);
}
}
return 0;
}
int mlx4_en_timestamp_config(struct net_device *dev, int tx_type, int rx_filter)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
struct mlx4_en_dev *mdev = priv->mdev;
int port_up = 0;
int err = 0;
mutex_lock(&mdev->state_lock);
if (priv->port_up) {
port_up = 1;
mlx4_en_stop_port(dev, 1);
}
mlx4_en_free_resources(priv);
en_warn(priv, "Changing Time Stamp configuration\n");
priv->hwtstamp_config.tx_type = tx_type;
priv->hwtstamp_config.rx_filter = rx_filter;
if (rx_filter != HWTSTAMP_FILTER_NONE)
dev->features &= ~NETIF_F_HW_VLAN_RX;
else
dev->features |= NETIF_F_HW_VLAN_RX;
err = mlx4_en_alloc_resources(priv);
if (err) {
en_err(priv, "Failed reallocating port resources\n");
goto out;
}
if (port_up) {
err = mlx4_en_start_port(dev);
if (err)
en_err(priv, "Failed starting port\n");
}
out:
mutex_unlock(&mdev->state_lock);
netdev_features_change(dev);
return err;
}
netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct skb_shared_info *shinfo = skb_shinfo(skb);
struct mlx4_en_priv *priv = netdev_priv(dev);
union mlx4_wqe_qpn_vlan qpn_vlan = {};
struct mlx4_en_tx_ring *ring;
struct mlx4_en_tx_desc *tx_desc;
struct mlx4_wqe_data_seg *data;
struct mlx4_en_tx_info *tx_info;
int tx_ind;
int nr_txbb;
int desc_size;
int real_size;
u32 index, bf_index;
__be32 op_own;
int lso_header_size;
void *fragptr = NULL;
bool bounce = false;
bool send_doorbell;
bool stop_queue;
bool inline_ok;
u8 data_offset;
u32 ring_cons;
bool bf_ok;
tx_ind = skb_get_queue_mapping(skb);
ring = priv->tx_ring[TX][tx_ind];
if (unlikely(!priv->port_up))
goto tx_drop;
/* fetch ring->cons far ahead before needing it to avoid stall */
ring_cons = READ_ONCE(ring->cons);
real_size = get_real_size(skb, shinfo, dev, &lso_header_size,
&inline_ok, &fragptr);
if (unlikely(!real_size))
goto tx_drop_count;
/* Align descriptor to TXBB size */
desc_size = ALIGN(real_size, TXBB_SIZE);
nr_txbb = desc_size >> LOG_TXBB_SIZE;
if (unlikely(nr_txbb > MAX_DESC_TXBBS)) {
if (netif_msg_tx_err(priv))
en_warn(priv, "Oversized header or SG list\n");
goto tx_drop_count;
}
bf_ok = ring->bf_enabled;
if (skb_vlan_tag_present(skb)) {
u16 vlan_proto;
qpn_vlan.vlan_tag = cpu_to_be16(skb_vlan_tag_get(skb));
vlan_proto = be16_to_cpu(skb->vlan_proto);
if (vlan_proto == ETH_P_8021AD)
qpn_vlan.ins_vlan = MLX4_WQE_CTRL_INS_SVLAN;
else if (vlan_proto == ETH_P_8021Q)
qpn_vlan.ins_vlan = MLX4_WQE_CTRL_INS_CVLAN;
else
qpn_vlan.ins_vlan = 0;
bf_ok = false;
}
netdev_txq_bql_enqueue_prefetchw(ring->tx_queue);
/* Track current inflight packets for performance analysis */
AVG_PERF_COUNTER(priv->pstats.inflight_avg,
(u32)(ring->prod - ring_cons - 1));
/* Packet is good - grab an index and transmit it */
index = ring->prod & ring->size_mask;
bf_index = ring->prod;
/* See if we have enough space for whole descriptor TXBB for setting
* SW ownership on next descriptor; if not, use a bounce buffer. */
if (likely(index + nr_txbb <= ring->size))
tx_desc = ring->buf + (index << LOG_TXBB_SIZE);
else {
tx_desc = (struct mlx4_en_tx_desc *) ring->bounce_buf;
bounce = true;
bf_ok = false;
}
/* Save skb in tx_info ring */
tx_info = &ring->tx_info[index];
tx_info->skb = skb;
tx_info->nr_txbb = nr_txbb;
if (!lso_header_size) {
data = &tx_desc->data;
data_offset = offsetof(struct mlx4_en_tx_desc, data);
} else {
u32 mlx4_en_recycle_tx_desc(struct mlx4_en_priv *priv,
struct mlx4_en_tx_ring *ring,
int index, u64 timestamp,
int napi_mode)
{
struct mlx4_en_tx_info *tx_info = &ring->tx_info[index];
struct mlx4_en_rx_alloc frame = {
.page = tx_info->page,
.dma = tx_info->map0_dma,
};
if (!mlx4_en_rx_recycle(ring->recycle_ring, &frame)) {
dma_unmap_page(priv->ddev, tx_info->map0_dma,
PAGE_SIZE, priv->dma_dir);
put_page(tx_info->page);
}
return tx_info->nr_txbb;
}
int mlx4_en_free_tx_buf(struct net_device *dev, struct mlx4_en_tx_ring *ring)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
int cnt = 0;
/* Skip last polled descriptor */
ring->cons += ring->last_nr_txbb;
en_dbg(DRV, priv, "Freeing Tx buf - cons:0x%x prod:0x%x\n",
ring->cons, ring->prod);
if ((u32) (ring->prod - ring->cons) > ring->size) {
if (netif_msg_tx_err(priv))
en_warn(priv, "Tx consumer passed producer!\n");
return 0;
}
while (ring->cons != ring->prod) {
ring->last_nr_txbb = ring->free_tx_desc(priv, ring,
ring->cons & ring->size_mask,
0, 0 /* Non-NAPI caller */);
ring->cons += ring->last_nr_txbb;
cnt++;
}
if (ring->tx_queue)
netdev_tx_reset_queue(ring->tx_queue);
if (cnt)
en_dbg(DRV, priv, "Freed %d uncompleted tx descriptors\n", cnt);
return cnt;
}
bool mlx4_en_process_tx_cq(struct net_device *dev,
struct mlx4_en_cq *cq, int napi_budget)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
struct mlx4_cq *mcq = &cq->mcq;
struct mlx4_en_tx_ring *ring = priv->tx_ring[cq->type][cq->ring];
struct mlx4_cqe *cqe;
u16 index, ring_index, stamp_index;
u32 txbbs_skipped = 0;
u32 txbbs_stamp = 0;
u32 cons_index = mcq->cons_index;
int size = cq->size;
u32 size_mask = ring->size_mask;
struct mlx4_cqe *buf = cq->buf;
u32 packets = 0;
u32 bytes = 0;
int factor = priv->cqe_factor;
int done = 0;
int budget = priv->tx_work_limit;
u32 last_nr_txbb;
u32 ring_cons;
if (unlikely(!priv->port_up))
return true;
netdev_txq_bql_complete_prefetchw(ring->tx_queue);
index = cons_index & size_mask;
cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor;
last_nr_txbb = READ_ONCE(ring->last_nr_txbb);
ring_cons = READ_ONCE(ring->cons);
ring_index = ring_cons & size_mask;
stamp_index = ring_index;
/* Process all completed CQEs */
while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK,
cons_index & size) && (done < budget)) {
u16 new_index;
/*
* make sure we read the CQE after we read the
* ownership bit
*/
dma_rmb();
if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) ==
MLX4_CQE_OPCODE_ERROR)) {
struct mlx4_err_cqe *cqe_err = (struct mlx4_err_cqe *)cqe;
en_err(priv, "CQE error - vendor syndrome: 0x%x syndrome: 0x%x\n",
cqe_err->vendor_err_syndrome,
cqe_err->syndrome);
}
/* Skip over last polled CQE */
new_index = be16_to_cpu(cqe->wqe_index) & size_mask;
do {
u64 timestamp = 0;
txbbs_skipped += last_nr_txbb;
ring_index = (ring_index + last_nr_txbb) & size_mask;
if (unlikely(ring->tx_info[ring_index].ts_requested))
timestamp = mlx4_en_get_cqe_ts(cqe);
/* free next descriptor */
last_nr_txbb = ring->free_tx_desc(
priv, ring, ring_index,
timestamp, napi_budget);
mlx4_en_stamp_wqe(priv, ring, stamp_index,
!!((ring_cons + txbbs_stamp) &
ring->size));
stamp_index = ring_index;
txbbs_stamp = txbbs_skipped;
packets++;
bytes += ring->tx_info[ring_index].nr_bytes;
} while ((++done < budget) && (ring_index != new_index));
++cons_index;
index = cons_index & size_mask;
cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor;
}
/*
* To prevent CQ overflow we first update CQ consumer and only then
* the ring consumer.
*/
mcq->cons_index = cons_index;
mlx4_cq_set_ci(mcq);
wmb();
/* we want to dirty this cache line once */
WRITE_ONCE(ring->last_nr_txbb, last_nr_txbb);
WRITE_ONCE(ring->cons, ring_cons + txbbs_skipped);
if (cq->type == TX_XDP)
return done < budget;
netdev_tx_completed_queue(ring->tx_queue, packets, bytes);
/* Wakeup Tx queue if this stopped, and ring is not full.
*/
if (netif_tx_queue_stopped(ring->tx_queue) &&
!mlx4_en_is_tx_ring_full(ring)) {
netif_tx_wake_queue(ring->tx_queue);
ring->wake_queue++;
}
return done < budget;
}
void mlx4_en_tx_irq(struct mlx4_cq *mcq)
{
struct mlx4_en_cq *cq = container_of(mcq, struct mlx4_en_cq, mcq);
struct mlx4_en_priv *priv = netdev_priv(cq->dev);
if (likely(priv->port_up))
napi_schedule_irqoff(&cq->napi);
else
mlx4_en_arm_cq(priv, cq);
}
/* TX CQ polling - called by NAPI */
int mlx4_en_poll_tx_cq(struct napi_struct *napi, int budget)
{
struct mlx4_en_cq *cq = container_of(napi, struct mlx4_en_cq, napi);
struct net_device *dev = cq->dev;
struct mlx4_en_priv *priv = netdev_priv(dev);
bool clean_complete;
clean_complete = mlx4_en_process_tx_cq(dev, cq, budget);
if (!clean_complete)
return budget;
napi_complete(napi);
mlx4_en_arm_cq(priv, cq);
return 0;
}
static struct mlx4_en_tx_desc *mlx4_en_bounce_to_desc(struct mlx4_en_priv *priv,
struct mlx4_en_tx_ring *ring,
u32 index,
unsigned int desc_size)
{
u32 copy = (ring->size - index) << LOG_TXBB_SIZE;
int i;
for (i = desc_size - copy - 4; i >= 0; i -= 4) {
if ((i & (TXBB_SIZE - 1)) == 0)
wmb();
*((u32 *) (ring->buf + i)) =
*((u32 *) (ring->bounce_buf + copy + i));
}
for (i = copy - 4; i >= 4 ; i -= 4) {
if ((i & (TXBB_SIZE - 1)) == 0)
wmb();
*((u32 *)(ring->buf + (index << LOG_TXBB_SIZE) + i)) =
*((u32 *) (ring->bounce_buf + i));
}
/* Return real descriptor location */
return ring->buf + (index << LOG_TXBB_SIZE);
}