int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int budget)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
struct mlx4_cqe *cqe;
struct mlx4_en_rx_ring *ring = &priv->rx_ring[cq->ring];
struct skb_frag_struct *skb_frags;
struct mlx4_en_rx_desc *rx_desc;
struct sk_buff *skb;
int index;
int nr;
unsigned int length;
int polled = 0;
int ip_summed;
if (!priv->port_up)
return 0;
/* We assume a 1:1 mapping between CQEs and Rx descriptors, so Rx
* descriptor offset can be deduced from the CQE index instead of
* reading 'cqe->index' */
index = cq->mcq.cons_index & ring->size_mask;
cqe = &cq->buf[index];
/* Process all completed CQEs */
while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK,
cq->mcq.cons_index & cq->size)) {
skb_frags = ring->rx_info + (index << priv->log_rx_info);
rx_desc = ring->buf + (index << ring->log_stride);
/*
* make sure we read the CQE after we read the ownership bit
*/
rmb();
/* Drop packet on bad receive or bad checksum */
if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) ==
MLX4_CQE_OPCODE_ERROR)) {
en_err(priv, "CQE completed in error - vendor "
"syndrom:%d syndrom:%d\n",
((struct mlx4_err_cqe *) cqe)->vendor_err_syndrome,
((struct mlx4_err_cqe *) cqe)->syndrome);
goto next;
}
if (unlikely(cqe->badfcs_enc & MLX4_CQE_BAD_FCS)) {
en_dbg(RX_ERR, priv, "Accepted frame with bad FCS\n");
goto next;
}
/*
* Packet is OK - process it.
*/
length = be32_to_cpu(cqe->byte_cnt);
ring->bytes += length;
ring->packets++;
if (likely(dev->features & NETIF_F_RXCSUM)) {
if ((cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPOK)) &&
(cqe->checksum == cpu_to_be16(0xffff))) {
priv->port_stats.rx_chksum_good++;
/* This packet is eligible for LRO if it is:
* - DIX Ethernet (type interpretation)
* - TCP/IP (v4)
* - without IP options
* - not an IP fragment */
if (dev->features & NETIF_F_GRO) {
struct sk_buff *gro_skb = napi_get_frags(&cq->napi);
if (!gro_skb)
goto next;
nr = mlx4_en_complete_rx_desc(
priv, rx_desc,
skb_frags, skb_shinfo(gro_skb)->frags,
ring->page_alloc, length);
if (!nr)
goto next;
skb_shinfo(gro_skb)->nr_frags = nr;
gro_skb->len = length;
gro_skb->data_len = length;
gro_skb->truesize += length;
gro_skb->ip_summed = CHECKSUM_UNNECESSARY;
if (priv->vlgrp && (cqe->vlan_my_qpn &
cpu_to_be32(MLX4_CQE_VLAN_PRESENT_MASK)))
vlan_gro_frags(&cq->napi, priv->vlgrp, be16_to_cpu(cqe->sl_vid));
else
napi_gro_frags(&cq->napi);
goto next;
}
/* LRO not possible, complete processing here */
ip_summed = CHECKSUM_UNNECESSARY;
} else {
ip_summed = CHECKSUM_NONE;
priv->port_stats.rx_chksum_none++;
}
} else {
ip_summed = CHECKSUM_NONE;
priv->port_stats.rx_chksum_none++;
}
skb = mlx4_en_rx_skb(priv, rx_desc, skb_frags,
ring->page_alloc, length);
if (!skb) {
priv->stats.rx_dropped++;
goto next;
}
if (unlikely(priv->validate_loopback)) {
validate_loopback(priv, skb);
goto next;
}
skb->ip_summed = ip_summed;
skb->protocol = eth_type_trans(skb, dev);
skb_record_rx_queue(skb, cq->ring);
/* Push it up the stack */
if (priv->vlgrp && (be32_to_cpu(cqe->vlan_my_qpn) &
MLX4_CQE_VLAN_PRESENT_MASK)) {
vlan_hwaccel_receive_skb(skb, priv->vlgrp,
be16_to_cpu(cqe->sl_vid));
} else
netif_receive_skb(skb);
next:
++cq->mcq.cons_index;
index = (cq->mcq.cons_index) & ring->size_mask;
cqe = &cq->buf[index];
if (++polled == budget) {
/* We are here because we reached the NAPI budget -
* flush only pending LRO sessions */
goto out;
}
}
out:
AVG_PERF_COUNTER(priv->pstats.rx_coal_avg, polled);
mlx4_cq_set_ci(&cq->mcq);
wmb(); /* ensure HW sees CQ consumer before we post new buffers */
ring->cons = cq->mcq.cons_index;
ring->prod += polled; /* Polled descriptors were realocated in place */
mlx4_en_update_rx_prod_db(ring);
return polled;
}
static void mlx4_en_process_tx_cq(struct net_device *dev, struct mlx4_en_cq *cq)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
struct mlx4_cq *mcq = &cq->mcq;
struct mlx4_en_tx_ring *ring = &priv->tx_ring[cq->ring];
struct mlx4_cqe *cqe = cq->buf;
u16 index;
u16 new_index;
u32 txbbs_skipped = 0;
u32 cq_last_sav;
/* index always points to the first TXBB of the last polled descriptor */
index = ring->cons & ring->size_mask;
new_index = be16_to_cpu(cqe->wqe_index) & ring->size_mask;
if (index == new_index)
return;
if (!priv->port_up)
return;
/*
* We use a two-stage loop:
* - the first samples the HW-updated CQE
* - the second frees TXBBs until the last sample
* This lets us amortize CQE cache misses, while still polling the CQ
* until is quiescent.
*/
cq_last_sav = mcq->cons_index;
do {
do {
/* Skip over last polled CQE */
index = (index + ring->last_nr_txbb) & ring->size_mask;
txbbs_skipped += ring->last_nr_txbb;
/* Poll next CQE */
ring->last_nr_txbb = mlx4_en_free_tx_desc(
priv, ring, index,
!!((ring->cons + txbbs_skipped) &
ring->size));
++mcq->cons_index;
} while (index != new_index);
new_index = be16_to_cpu(cqe->wqe_index) & ring->size_mask;
} while (index != new_index);
AVG_PERF_COUNTER(priv->pstats.tx_coal_avg,
(u32) (mcq->cons_index - cq_last_sav));
/*
* To prevent CQ overflow we first update CQ consumer and only then
* the ring consumer.
*/
mlx4_cq_set_ci(mcq);
wmb();
ring->cons += txbbs_skipped;
/* Wakeup Tx queue if this ring stopped it */
if (unlikely(ring->blocked)) {
if ((u32) (ring->prod - ring->cons) <=
ring->size - HEADROOM - MAX_DESC_TXBBS) {
ring->blocked = 0;
netif_tx_wake_queue(netdev_get_tx_queue(dev, cq->ring));
priv->port_stats.wake_queue++;
}
}
}
int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int budget)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
struct mlx4_cqe *cqe;
struct mlx4_en_rx_ring *ring = &priv->rx_ring[cq->ring];
struct mbuf **mb_list;
struct mlx4_en_rx_desc *rx_desc;
struct mbuf *mb;
#ifdef INET
struct lro_entry *queued;
#endif
int index;
unsigned int length;
int polled = 0;
if (!priv->port_up)
return 0;
/* We assume a 1:1 mapping between CQEs and Rx descriptors, so Rx
* descriptor offset can be deduced from the CQE index instead of
* reading 'cqe->index' */
index = cq->mcq.cons_index & ring->size_mask;
cqe = &cq->buf[index];
/* Process all completed CQEs */
while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK,
cq->mcq.cons_index & cq->size)) {
mb_list = ring->rx_info + (index << priv->log_rx_info);
rx_desc = ring->buf + (index << ring->log_stride);
/*
* make sure we read the CQE after we read the ownership bit
*/
rmb();
if (invalid_cqe(priv, cqe))
goto next;
/*
* Packet is OK - process it.
*/
length = be32_to_cpu(cqe->byte_cnt);
mb = mlx4_en_rx_mb(priv, rx_desc, mb_list, length);
if (!mb) {
ring->errors++;
goto next;
}
ring->bytes += length;
ring->packets++;
if (unlikely(priv->validate_loopback)) {
validate_loopback(priv, mb);
goto next;
}
mb->m_pkthdr.flowid = cq->ring;
mb->m_flags |= M_FLOWID;
mb->m_pkthdr.rcvif = dev;
if (be32_to_cpu(cqe->vlan_my_qpn) &
MLX4_CQE_VLAN_PRESENT_MASK) {
mb->m_pkthdr.ether_vtag = be16_to_cpu(cqe->sl_vid);
mb->m_flags |= M_VLANTAG;
}
if (likely(priv->rx_csum) &&
(cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPOK)) &&
(cqe->checksum == cpu_to_be16(0xffff))) {
priv->port_stats.rx_chksum_good++;
mb->m_pkthdr.csum_flags =
CSUM_IP_CHECKED | CSUM_IP_VALID |
CSUM_DATA_VALID | CSUM_PSEUDO_HDR;
mb->m_pkthdr.csum_data = htons(0xffff);
/* This packet is eligible for LRO if it is:
* - DIX Ethernet (type interpretation)
* - TCP/IP (v4)
* - without IP options
* - not an IP fragment
*/
#ifdef INET
if (mlx4_en_can_lro(cqe->status) &&
(dev->if_capenable & IFCAP_LRO)) {
if (ring->lro.lro_cnt != 0 &&
tcp_lro_rx(&ring->lro, mb, 0) == 0)
goto next;
}
#endif
/* LRO not possible, complete processing here */
INC_PERF_COUNTER(priv->pstats.lro_misses);
} else {
mb->m_pkthdr.csum_flags = 0;
priv->port_stats.rx_chksum_none++;
#ifdef INET
if (priv->ip_reasm &&
cqe->status & cpu_to_be16(MLX4_CQE_STATUS_IPV4) &&
!mlx4_en_rx_frags(priv, ring, mb, cqe))
goto next;
#endif
}
/* Push it up the stack */
dev->if_input(dev, mb);
next:
++cq->mcq.cons_index;
index = (cq->mcq.cons_index) & ring->size_mask;
cqe = &cq->buf[index];
if (++polled == budget)
goto out;
}
/* Flush all pending IP reassembly sessions */
out:
#ifdef INET
mlx4_en_flush_frags(priv, ring);
while ((queued = SLIST_FIRST(&ring->lro.lro_active)) != NULL) {
SLIST_REMOVE_HEAD(&ring->lro.lro_active, next);
tcp_lro_flush(&ring->lro, queued);
}
#endif
AVG_PERF_COUNTER(priv->pstats.rx_coal_avg, polled);
mlx4_cq_set_ci(&cq->mcq);
wmb(); /* ensure HW sees CQ consumer before we post new buffers */
ring->cons = cq->mcq.cons_index;
ring->prod += polled; /* Polled descriptors were realocated in place */
mlx4_en_update_rx_prod_db(ring);
return polled;
}
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 {
int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int budget)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
struct mlx4_en_dev *mdev = priv->mdev;
struct mlx4_cqe *cqe;
struct mlx4_en_rx_ring *ring = &priv->rx_ring[cq->ring];
struct skb_frag_struct *skb_frags;
struct mlx4_en_rx_desc *rx_desc;
struct sk_buff *skb;
int index;
unsigned int length;
int polled = 0;
int ip_summed;
if (!priv->port_up)
return 0;
/* We assume a 1:1 mapping between CQEs and Rx descriptors, so Rx
* descriptor offset can be deduced from the CQE index instead of
* reading 'cqe->index' */
index = cq->mcq.cons_index & ring->size_mask;
cqe = &cq->buf[index];
/* Process all completed CQEs */
while (XNOR(cqe->owner_sr_opcode & MLX4_CQE_OWNER_MASK,
cq->mcq.cons_index & cq->size)) {
skb_frags = ring->rx_info + (index << priv->log_rx_info);
rx_desc = ring->buf + (index << ring->log_stride);
/*
* make sure we read the CQE after we read the ownership bit
*/
rmb();
/* Drop packet on bad receive or bad checksum */
if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) ==
MLX4_CQE_OPCODE_ERROR)) {
mlx4_err(mdev, "CQE completed in error - vendor "
"syndrom:%d syndrom:%d\n",
((struct mlx4_err_cqe *) cqe)->vendor_err_syndrome,
((struct mlx4_err_cqe *) cqe)->syndrome);
goto next;
}
if (unlikely(cqe->badfcs_enc & MLX4_CQE_BAD_FCS)) {
mlx4_dbg(RX_ERR, priv, "Accepted frame with bad FCS\n");
goto next;
}
/*
* Packet is OK - process it.
*/
length = be32_to_cpu(cqe->byte_cnt);
ring->bytes += length;
ring->packets++;
if (likely(priv->rx_csum)) {
if ((cqe->status & MLX4_CQE_STATUS_IPOK) &&
(cqe->checksum == 0xffff)) {
priv->port_stats.rx_chksum_good++;
if (mdev->profile.num_lro &&
!mlx4_en_lro_rx(priv, ring, rx_desc,
skb_frags, length, cqe))
goto next;
/* LRO not possible, complete processing here */
ip_summed = CHECKSUM_UNNECESSARY;
INC_PERF_COUNTER(priv->pstats.lro_misses);
} else {
ip_summed = CHECKSUM_NONE;
priv->port_stats.rx_chksum_none++;
}
} else {
ip_summed = CHECKSUM_NONE;
priv->port_stats.rx_chksum_none++;
}
skb = mlx4_en_rx_skb(priv, rx_desc, skb_frags,
ring->page_alloc, length);
if (!skb) {
priv->stats.rx_dropped++;
goto next;
}
skb->ip_summed = ip_summed;
skb->protocol = eth_type_trans(skb, dev);
/* Push it up the stack */
if (priv->vlgrp && (be32_to_cpu(cqe->vlan_my_qpn) &
MLX4_CQE_VLAN_PRESENT_MASK)) {
vlan_hwaccel_receive_skb(skb, priv->vlgrp,
be16_to_cpu(cqe->sl_vid));
} else
netif_receive_skb(skb);
dev->last_rx = jiffies;
next:
++cq->mcq.cons_index;
index = (cq->mcq.cons_index) & ring->size_mask;
cqe = &cq->buf[index];
if (++polled == budget) {
/* We are here because we reached the NAPI budget -
* flush only pending LRO sessions */
if (mdev->profile.num_lro)
mlx4_en_lro_flush(priv, ring, 0);
goto out;
}
}
/* If CQ is empty flush all LRO sessions unconditionally */
if (mdev->profile.num_lro)
mlx4_en_lro_flush(priv, ring, 1);
out:
AVG_PERF_COUNTER(priv->pstats.rx_coal_avg, polled);
mlx4_cq_set_ci(&cq->mcq);
wmb(); /* ensure HW sees CQ consumer before we post new buffers */
ring->cons = cq->mcq.cons_index;
ring->prod += polled; /* Polled descriptors were realocated in place */
if (unlikely(!ring->full)) {
mlx4_en_copy_desc(priv, ring, ring->cons - polled,
ring->prod - polled, polled);
mlx4_en_fill_rx_buf(dev, ring);
}
mlx4_en_update_rx_prod_db(ring);
return polled;
}
