Skip to content

Commit

Permalink
mlx4_en: Not using Shared Receive Queues
Browse files Browse the repository at this point in the history 
We use 1:1 mapping between QPs and SRQs on receive side,
so additional indirection level not required. Allocated the receive
buffers for the RSS QPs.

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
  • Loading branch information
Yevgeny Petrilin authored and David S. Miller committed Aug 7, 2009
1 parent b6b912e commit 9f519f6
Show file tree
Hide file tree
Showing 5 changed files with 26 additions and 73 deletions.
3 changes: 1 addition & 2 deletions drivers/net/mlx4/en_netdev.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -622,8 +622,7 @@ int mlx4_en_start_port(struct net_device *dev)

/* Configure ring */
tx_ring = &priv->tx_ring[i];
err = mlx4_en_activate_tx_ring(priv, tx_ring, cq->mcq.cqn,
priv->rx_ring[0].srq.srqn);
err = mlx4_en_activate_tx_ring(priv, tx_ring, cq->mcq.cqn);
if (err) {
en_err(priv, "Failed allocating Tx ring\n");
mlx4_en_deactivate_cq(priv, cq);
Expand Down
9 changes: 4 additions & 5 deletions drivers/net/mlx4/en_resources.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -37,7 +37,7 @@
#include "mlx4_en.h"

void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
int is_tx, int rss, int qpn, int cqn, int srqn,
int is_tx, int rss, int qpn, int cqn,
struct mlx4_qp_context *context)
{
struct mlx4_en_dev *mdev = priv->mdev;
Expand All @@ -46,11 +46,12 @@ void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
context->flags = cpu_to_be32(7 << 16 | rss << 13);
context->pd = cpu_to_be32(mdev->priv_pdn);
context->mtu_msgmax = 0xff;
context->rq_size_stride = 0;
if (!is_tx && !rss)
context->rq_size_stride = ilog2(size) << 3 | (ilog2(stride) - 4);
if (is_tx)
context->sq_size_stride = ilog2(size) << 3 | (ilog2(stride) - 4);
else
context->sq_size_stride = 1;
context->sq_size_stride = ilog2(TXBB_SIZE) - 4;
context->usr_page = cpu_to_be32(mdev->priv_uar.index);
context->local_qpn = cpu_to_be32(qpn);
context->pri_path.ackto = 1 & 0x07;
Expand All @@ -59,8 +60,6 @@ void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
context->cqn_send = cpu_to_be32(cqn);
context->cqn_recv = cpu_to_be32(cqn);
context->db_rec_addr = cpu_to_be64(priv->res.db.dma << 2);
if (!rss)
context->srqn = cpu_to_be32(MLX4_EN_USE_SRQ | srqn);
}


Expand Down
77 changes: 17 additions & 60 deletions drivers/net/mlx4/en_rx.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -40,16 +40,6 @@

#include "mlx4_en.h"

static void *get_wqe(struct mlx4_en_rx_ring *ring, int n)
{
int offset = n << ring->srq.wqe_shift;
return ring->buf + offset;
}

static void mlx4_en_srq_event(struct mlx4_srq *srq, enum mlx4_event type)
{
return;
}

static int mlx4_en_get_frag_header(struct skb_frag_struct *frags, void **mac_hdr,
void **ip_hdr, void **tcpudp_hdr,
Expand Down Expand Up @@ -154,9 +144,6 @@ static void mlx4_en_init_rx_desc(struct mlx4_en_priv *priv,
int possible_frags;
int i;

/* Pre-link descriptor */
rx_desc->next.next_wqe_index = cpu_to_be16((index + 1) & ring->size_mask);

/* Set size and memtype fields */
for (i = 0; i < priv->num_frags; i++) {
skb_frags[i].size = priv->frag_info[i].frag_size;
Expand Down Expand Up @@ -294,17 +281,14 @@ int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv,
int err;
int tmp;

/* Sanity check SRQ size before proceeding */
if (size >= mdev->dev->caps.max_srq_wqes)
return -EINVAL;

ring->prod = 0;
ring->cons = 0;
ring->size = size;
ring->size_mask = size - 1;
ring->stride = stride;
ring->log_stride = ffs(ring->stride) - 1;
ring->buf_size = ring->size * ring->stride;
ring->buf_size = ring->size * ring->stride + TXBB_SIZE;

tmp = size * roundup_pow_of_two(MLX4_EN_MAX_RX_FRAGS *
sizeof(struct skb_frag_struct));
Expand Down Expand Up @@ -360,15 +344,12 @@ int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv,

int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv)
{
struct mlx4_en_dev *mdev = priv->mdev;
struct mlx4_wqe_srq_next_seg *next;
struct mlx4_en_rx_ring *ring;
int i;
int ring_ind;
int err;
int stride = roundup_pow_of_two(sizeof(struct mlx4_en_rx_desc) +
DS_SIZE * priv->num_frags);
int max_gs = (stride - sizeof(struct mlx4_wqe_srq_next_seg)) / DS_SIZE;

for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++) {
ring = &priv->rx_ring[ring_ind];
Expand All @@ -379,6 +360,9 @@ int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv)
ring->cqn = priv->rx_cq[ring_ind].mcq.cqn;

ring->stride = stride;
if (ring->stride <= TXBB_SIZE)
ring->buf += TXBB_SIZE;

ring->log_stride = ffs(ring->stride) - 1;
ring->buf_size = ring->size * ring->stride;

Expand All @@ -405,37 +389,10 @@ int mlx4_en_activate_rx_rings(struct mlx4_en_priv *priv)
ring = &priv->rx_ring[ring_ind];

mlx4_en_update_rx_prod_db(ring);

/* Configure SRQ representing the ring */
ring->srq.max = ring->actual_size;
ring->srq.max_gs = max_gs;
ring->srq.wqe_shift = ilog2(ring->stride);

for (i = 0; i < ring->srq.max; ++i) {
next = get_wqe(ring, i);
next->next_wqe_index =
cpu_to_be16((i + 1) & (ring->srq.max - 1));
}

err = mlx4_srq_alloc(mdev->dev, mdev->priv_pdn, &ring->wqres.mtt,
ring->wqres.db.dma, &ring->srq);
if (err){
en_err(priv, "Failed to allocate srq\n");
ring_ind--;
goto err_srq;
}
ring->srq.event = mlx4_en_srq_event;
}

return 0;

err_srq:
while (ring_ind >= 0) {
ring = &priv->rx_ring[ring_ind];
mlx4_srq_free(mdev->dev, &ring->srq);
ring_ind--;
}

err_buffers:
for (ring_ind = 0; ring_ind < priv->rx_ring_num; ring_ind++)
mlx4_en_free_rx_buf(priv, &priv->rx_ring[ring_ind]);
Expand All @@ -456,18 +413,17 @@ void mlx4_en_destroy_rx_ring(struct mlx4_en_priv *priv,

kfree(ring->lro.lro_arr);
mlx4_en_unmap_buffer(&ring->wqres.buf);
mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size);
mlx4_free_hwq_res(mdev->dev, &ring->wqres, ring->buf_size + TXBB_SIZE);
vfree(ring->rx_info);
ring->rx_info = NULL;
}

void mlx4_en_deactivate_rx_ring(struct mlx4_en_priv *priv,
struct mlx4_en_rx_ring *ring)
{
struct mlx4_en_dev *mdev = priv->mdev;

mlx4_srq_free(mdev->dev, &ring->srq);
mlx4_en_free_rx_buf(priv, ring);
if (ring->stride <= TXBB_SIZE)
ring->buf -= TXBB_SIZE;
mlx4_en_destroy_allocator(priv, ring);
}

Expand Down Expand Up @@ -835,8 +791,8 @@ void mlx4_en_calc_rx_buf(struct net_device *dev)

/* RSS related functions */

static int mlx4_en_config_rss_qp(struct mlx4_en_priv *priv,
int qpn, int srqn, int cqn,
static int mlx4_en_config_rss_qp(struct mlx4_en_priv *priv, int qpn,
struct mlx4_en_rx_ring *ring,
enum mlx4_qp_state *state,
struct mlx4_qp *qp)
{
Expand All @@ -858,13 +814,16 @@ static int mlx4_en_config_rss_qp(struct mlx4_en_priv *priv,
qp->event = mlx4_en_sqp_event;

memset(context, 0, sizeof *context);
mlx4_en_fill_qp_context(priv, 0, 0, 0, 0, qpn, cqn, srqn, context);
mlx4_en_fill_qp_context(priv, ring->size, ring->stride, 0, 0,
qpn, ring->cqn, context);
context->db_rec_addr = cpu_to_be64(ring->wqres.db.dma);

err = mlx4_qp_to_ready(mdev->dev, &priv->res.mtt, context, qp, state);
err = mlx4_qp_to_ready(mdev->dev, &ring->wqres.mtt, context, qp, state);
if (err) {
mlx4_qp_remove(mdev->dev, qp);
mlx4_qp_free(mdev->dev, qp);
}
mlx4_en_update_rx_prod_db(ring);
out:
kfree(context);
return err;
Expand All @@ -880,7 +839,7 @@ int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv)
void *ptr;
int rss_xor = mdev->profile.rss_xor;
u8 rss_mask = mdev->profile.rss_mask;
int i, srqn, qpn, cqn;
int i, qpn;
int err = 0;
int good_qps = 0;

Expand All @@ -894,10 +853,8 @@ int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv)
}

for (i = 0; i < priv->rx_ring_num; i++) {
cqn = priv->rx_ring[i].cqn;
srqn = priv->rx_ring[i].srq.srqn;
qpn = rss_map->base_qpn + i;
err = mlx4_en_config_rss_qp(priv, qpn, srqn, cqn,
err = mlx4_en_config_rss_qp(priv, qpn, &priv->rx_ring[i],
&rss_map->state[i],
&rss_map->qps[i]);
if (err)
Expand All @@ -920,7 +877,7 @@ int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv)
}
rss_map->indir_qp.event = mlx4_en_sqp_event;
mlx4_en_fill_qp_context(priv, 0, 0, 0, 1, priv->base_qpn,
priv->rx_ring[0].cqn, 0, &context);
priv->rx_ring[0].cqn, &context);

ptr = ((void *) &context) + 0x3c;
rss_context = (struct mlx4_en_rss_context *) ptr;
Expand Down
4 changes: 2 additions & 2 deletions drivers/net/mlx4/en_tx.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -150,7 +150,7 @@ void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv,

int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
struct mlx4_en_tx_ring *ring,
int cq, int srqn)
int cq)
{
struct mlx4_en_dev *mdev = priv->mdev;
int err;
Expand All @@ -168,7 +168,7 @@ int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
ring->doorbell_qpn = swab32(ring->qp.qpn << 8);

mlx4_en_fill_qp_context(priv, ring->size, ring->stride, 1, 0, ring->qpn,
ring->cqn, srqn, &ring->context);
ring->cqn, &ring->context);

err = mlx4_qp_to_ready(mdev->dev, &ring->wqres.mtt, &ring->context,
&ring->qp, &ring->qp_state);
Expand Down
6 changes: 2 additions & 4 deletions drivers/net/mlx4/mlx4_en.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -274,13 +274,11 @@ struct mlx4_en_tx_ring {
};

struct mlx4_en_rx_desc {
struct mlx4_wqe_srq_next_seg next;
/* actual number of entries depends on rx ring stride */
struct mlx4_wqe_data_seg data[0];
};

struct mlx4_en_rx_ring {
struct mlx4_srq srq;
struct mlx4_hwq_resources wqres;
struct mlx4_en_rx_alloc page_alloc[MLX4_EN_MAX_RX_FRAGS];
struct net_lro_mgr lro;
Expand Down Expand Up @@ -527,7 +525,7 @@ int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ri
void mlx4_en_destroy_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring);
int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
struct mlx4_en_tx_ring *ring,
int cq, int srqn);
int cq);
void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv,
struct mlx4_en_tx_ring *ring);

Expand All @@ -544,7 +542,7 @@ int mlx4_en_process_rx_cq(struct net_device *dev,
int budget);
int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget);
void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
int is_tx, int rss, int qpn, int cqn, int srqn,
int is_tx, int rss, int qpn, int cqn,
struct mlx4_qp_context *context);
void mlx4_en_sqp_event(struct mlx4_qp *qp, enum mlx4_event event);
int mlx4_en_map_buffer(struct mlx4_buf *buf);
Expand Down

0 comments on commit 9f519f6

Please sign in to comment.