Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-11-10 22:21:40 +00:00
Code Issues Packages Projects Releases Wiki Activity

net/mlx5e: Avoid TX CQE generation if more xmit packets expected

Browse Source 
In order to save PCI BW consumed by TX CQEs and to reduce the amount of
CPU cache misses caused by TX CQE reading, we request TX CQE generation
only when skb->xmit_more=0.

As a consequence of the above, a single TX CQE may now indicate the
transmission completion of multiple TX SKBs.

This also handles a problem introduced in commit b1b8105ebf41 "net/mlx5e:
Support NETIF_F_SG" where we didn't ask for NOP completions while the
driver didn't have the proper code to handle this case.

Fixes: b1b8105ebf41 ('net/mlx5e: Support NETIF_F_SG')
Signed-off-by: Achiad Shochat <achiad@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Achiad Shochat 2015-06-23 17:14:18 +03:00 committed by David S. Miller
parent 9fc5930625
commit 059ba072eb
1 changed files with 34 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

57
drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
View File

@ -210,7 +210,6 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | opcode);
cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt);
cseg->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
sq->skb[pi] = skb;
@ -225,8 +224,10 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
sq->stats.stopped++;
}
if (!skb->xmit_more || netif_xmit_stopped(sq->txq))
if (!skb->xmit_more || netif_xmit_stopped(sq->txq)) {
cseg->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
mlx5e_tx_notify_hw(sq, wqe);
}
/* fill sq edge with nops to avoid wqe wrap around */
while ((sq->pc & wq->sz_m1) > sq->edge)
@ -280,36 +281,46 @@ bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq)
for (i = 0; i < MLX5E_TX_CQ_POLL_BUDGET; i++) {
struct mlx5_cqe64 *cqe;
struct sk_buff *skb;
u16 ci;
int j;
u16 wqe_counter;
bool last_wqe;
cqe = mlx5e_get_cqe(cq);
if (!cqe)
break;
ci = sqcc & sq->wq.sz_m1;
skb = sq->skb[ci];
wqe_counter = be16_to_cpu(cqe->wqe_counter);
if (unlikely(!skb)) { /* nop */
sq->stats.nop++;
sqcc++;
continue;
}
do {
struct sk_buff *skb;
u16 ci;
int j;
for (j = 0; j < MLX5E_TX_SKB_CB(skb)->num_dma; j++) {
dma_addr_t addr;
u32 size;
last_wqe = (sqcc == wqe_counter);
mlx5e_dma_get(sq, dma_fifo_cc, &addr, &size);
dma_fifo_cc++;
dma_unmap_single(sq->pdev, addr, size, DMA_TO_DEVICE);
}
ci = sqcc & sq->wq.sz_m1;
skb = sq->skb[ci];
npkts++;
nbytes += MLX5E_TX_SKB_CB(skb)->num_bytes;
sqcc += MLX5E_TX_SKB_CB(skb)->num_wqebbs;
dev_kfree_skb(skb);
if (unlikely(!skb)) { /* nop */
sq->stats.nop++;
sqcc++;
continue;
}
for (j = 0; j < MLX5E_TX_SKB_CB(skb)->num_dma; j++) {
dma_addr_t addr;
u32 size;
mlx5e_dma_get(sq, dma_fifo_cc, &addr, &size);
dma_fifo_cc++;
dma_unmap_single(sq->pdev, addr, size,
DMA_TO_DEVICE);
}
npkts++;
nbytes += MLX5E_TX_SKB_CB(skb)->num_bytes;
sqcc += MLX5E_TX_SKB_CB(skb)->num_wqebbs;
dev_kfree_skb(skb);
} while (!last_wqe);
}
mlx5_cqwq_update_db_record(&cq->wq);