/**
* nfp_net_tx_tso() - Set up Tx descriptor for LSO
* @nn: NFP Net device
* @r_vec: per-ring structure
* @txbuf: Pointer to driver soft TX descriptor
* @txd: Pointer to HW TX descriptor
* @skb: Pointer to SKB
*
* 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 *nn, struct nfp_net_r_vector *r_vec,
struct nfp_net_tx_buf *txbuf,
struct nfp_net_tx_desc *txd, struct sk_buff *skb)
{
u32 hdrlen;
u16 mss;
if (!skb_is_gso(skb))
return;
if (!skb->encapsulation)
hdrlen = skb_transport_offset(skb) + tcp_hdrlen(skb);
else
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->l4_offset = hdrlen;
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);
}
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 netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
{
struct mlx5_wq_cyc *wq = &sq->wq;
u16 pi = sq->pc & wq->sz_m1;
struct mlx5e_tx_wqe *wqe = mlx5_wq_cyc_get_wqe(wq, pi);
struct mlx5e_tx_wqe_info *wi = &sq->wqe_info[pi];
struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
struct mlx5_wqe_eth_seg *eseg = &wqe->eth;
struct mlx5_wqe_data_seg *dseg;
unsigned char *skb_data = skb->data;
unsigned int skb_len = skb->len;
u8 opcode = MLX5_OPCODE_SEND;
dma_addr_t dma_addr = 0;
unsigned int num_bytes;
bool bf = false;
u16 headlen;
u16 ds_cnt = sizeof(*wqe) / MLX5_SEND_WQE_DS;
u16 ihs;
int i;
memset(wqe, 0, sizeof(*wqe));
if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
eseg->cs_flags = MLX5_ETH_WQE_L3_CSUM;
if (skb->encapsulation) {
eseg->cs_flags |= MLX5_ETH_WQE_L3_INNER_CSUM |
MLX5_ETH_WQE_L4_INNER_CSUM;
sq->stats.csum_partial_inner++;
} else {
eseg->cs_flags |= MLX5_ETH_WQE_L4_CSUM;
sq->stats.csum_partial++;
}
} else
sq->stats.csum_none++;
if (sq->cc != sq->prev_cc) {
sq->prev_cc = sq->cc;
sq->bf_budget = (sq->cc == sq->pc) ? MLX5E_SQ_BF_BUDGET : 0;
}
if (skb_is_gso(skb)) {
eseg->mss = cpu_to_be16(skb_shinfo(skb)->gso_size);
opcode = MLX5_OPCODE_LSO;
if (skb->encapsulation) {
ihs = skb_inner_transport_header(skb) - skb->data +
inner_tcp_hdrlen(skb);
sq->stats.tso_inner_packets++;
sq->stats.tso_inner_bytes += skb->len - ihs;
} else {
ihs = skb_transport_offset(skb) + tcp_hdrlen(skb);
sq->stats.tso_packets++;
sq->stats.tso_bytes += skb->len - ihs;
}
num_bytes = skb->len + (skb_shinfo(skb)->gso_segs - 1) * ihs;
} else {
bf = sq->bf_budget &&
!skb->xmit_more &&
!skb_shinfo(skb)->nr_frags;
ihs = mlx5e_get_inline_hdr_size(sq, skb, bf);
num_bytes = max_t(unsigned int, skb->len, ETH_ZLEN);
}
wi->num_bytes = num_bytes;
if (skb_vlan_tag_present(skb)) {
mlx5e_insert_vlan(eseg->inline_hdr_start, skb, ihs, &skb_data,
&skb_len);
ihs += VLAN_HLEN;
} else {
memcpy(eseg->inline_hdr_start, skb_data, ihs);
mlx5e_tx_skb_pull_inline(&skb_data, &skb_len, ihs);
}
eseg->inline_hdr_sz = cpu_to_be16(ihs);
ds_cnt += DIV_ROUND_UP(ihs - sizeof(eseg->inline_hdr_start),
MLX5_SEND_WQE_DS);
dseg = (struct mlx5_wqe_data_seg *)cseg + ds_cnt;
wi->num_dma = 0;
headlen = skb_len - skb->data_len;
if (headlen) {
dma_addr = dma_map_single(sq->pdev, skb_data, headlen,
DMA_TO_DEVICE);
if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
goto dma_unmap_wqe_err;
dseg->addr = cpu_to_be64(dma_addr);
dseg->lkey = sq->mkey_be;
dseg->byte_count = cpu_to_be32(headlen);
mlx5e_dma_push(sq, dma_addr, headlen, MLX5E_DMA_MAP_SINGLE);
wi->num_dma++;
dseg++;
}
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i];
int fsz = skb_frag_size(frag);
dma_addr = skb_frag_dma_map(sq->pdev, frag, 0, fsz,
DMA_TO_DEVICE);
if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
goto dma_unmap_wqe_err;
dseg->addr = cpu_to_be64(dma_addr);
dseg->lkey = sq->mkey_be;
dseg->byte_count = cpu_to_be32(fsz);
mlx5e_dma_push(sq, dma_addr, fsz, MLX5E_DMA_MAP_PAGE);
wi->num_dma++;
dseg++;
}
ds_cnt += wi->num_dma;
cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | opcode);
cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt);
sq->skb[pi] = skb;
wi->num_wqebbs = DIV_ROUND_UP(ds_cnt, MLX5_SEND_WQEBB_NUM_DS);
sq->pc += wi->num_wqebbs;
if (unlikely(MLX5E_TX_HW_STAMP(sq->channel->priv, skb)))
skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
netdev_tx_sent_queue(sq->txq, wi->num_bytes);
if (unlikely(!mlx5e_sq_has_room_for(sq, MLX5E_SQ_STOP_ROOM))) {
netif_tx_stop_queue(sq->txq);
sq->stats.queue_stopped++;
}
if (!skb->xmit_more || netif_xmit_stopped(sq->txq)) {
int bf_sz = 0;
if (bf && sq->uar_bf_map)
bf_sz = wi->num_wqebbs << 3;
cseg->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
mlx5e_tx_notify_hw(sq, &wqe->ctrl, bf_sz);
}
sq->bf_budget = bf ? sq->bf_budget - 1 : 0;
/* fill sq edge with nops to avoid wqe wrap around */
while ((sq->pc & wq->sz_m1) > sq->edge)
mlx5e_send_nop(sq, false);
sq->stats.packets++;
sq->stats.bytes += num_bytes;
return NETDEV_TX_OK;
dma_unmap_wqe_err:
sq->stats.queue_dropped++;
mlx5e_dma_unmap_wqe_err(sq, wi->num_dma);
dev_kfree_skb_any(skb);
return NETDEV_TX_OK;
}
