提交 811c0830 编写于 作者: Y Yunsheng Lin 提交者: David S. Miller

net: hns3: add tx send size handling for tso skb

The actual size on wire for tso skb should be (gso_segs - 1) *
hdr + skb->len instead of skb->len, which can be seen by user using
'ethtool -S ethX' cmd, and 'Byte Queue Limit' also use the send size
stat to do the queue limiting, so add send_bytes in the desc_cb to
record the actual send size for a skb. And send_bytes is only for tx
desc_cb and page_offset is only for rx desc, so reuse the same space
for both of them.
Signed-off-by: NYunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: NHuazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
上级 d5d5e019
......@@ -694,7 +694,7 @@ void hns3_enable_vlan_filter(struct net_device *netdev, bool enable)
}
static int hns3_set_tso(struct sk_buff *skb, u32 *paylen_fdop_ol4cs,
u16 *mss, u32 *type_cs_vlan_tso)
u16 *mss, u32 *type_cs_vlan_tso, u32 *send_bytes)
{
u32 l4_offset, hdr_len;
union l3_hdr_info l3;
......@@ -750,6 +750,8 @@ static int hns3_set_tso(struct sk_buff *skb, u32 *paylen_fdop_ol4cs,
(__force __wsum)htonl(l4_paylen));
}
*send_bytes = (skb_shinfo(skb)->gso_segs - 1) * hdr_len + skb->len;
/* find the txbd field values */
*paylen_fdop_ol4cs = skb->len - hdr_len;
hns3_set_field(*type_cs_vlan_tso, HNS3_TXD_TSO_B, 1);
......@@ -1076,7 +1078,8 @@ static bool hns3_check_hw_tx_csum(struct sk_buff *skb)
}
static int hns3_fill_skb_desc(struct hns3_enet_ring *ring,
struct sk_buff *skb, struct hns3_desc *desc)
struct sk_buff *skb, struct hns3_desc *desc,
struct hns3_desc_cb *desc_cb)
{
u32 ol_type_vlan_len_msec = 0;
u32 paylen_ol4cs = skb->len;
......@@ -1105,6 +1108,8 @@ static int hns3_fill_skb_desc(struct hns3_enet_ring *ring,
1);
}
desc_cb->send_bytes = skb->len;
if (skb->ip_summed == CHECKSUM_PARTIAL) {
u8 ol4_proto, il4_proto;
......@@ -1140,7 +1145,7 @@ static int hns3_fill_skb_desc(struct hns3_enet_ring *ring,
}
ret = hns3_set_tso(skb, &paylen_ol4cs, &mss_hw_csum,
&type_cs_vlan_tso);
&type_cs_vlan_tso, &desc_cb->send_bytes);
if (unlikely(ret < 0)) {
u64_stats_update_begin(&ring->syncp);
ring->stats.tx_tso_err++;
......@@ -1553,6 +1558,7 @@ netdev_tx_t hns3_nic_net_xmit(struct sk_buff *skb, struct net_device *netdev)
{
struct hns3_nic_priv *priv = netdev_priv(netdev);
struct hns3_enet_ring *ring = &priv->ring[skb->queue_mapping];
struct hns3_desc_cb *desc_cb = &ring->desc_cb[ring->next_to_use];
struct netdev_queue *dev_queue;
int pre_ntu, next_to_use_head;
bool doorbell;
......@@ -1580,7 +1586,8 @@ netdev_tx_t hns3_nic_net_xmit(struct sk_buff *skb, struct net_device *netdev)
next_to_use_head = ring->next_to_use;
ret = hns3_fill_skb_desc(ring, skb, &ring->desc[ring->next_to_use]);
ret = hns3_fill_skb_desc(ring, skb, &ring->desc[ring->next_to_use],
desc_cb);
if (unlikely(ret < 0))
goto fill_err;
......@@ -1600,7 +1607,7 @@ netdev_tx_t hns3_nic_net_xmit(struct sk_buff *skb, struct net_device *netdev)
/* Complete translate all packets */
dev_queue = netdev_get_tx_queue(netdev, ring->queue_index);
doorbell = __netdev_tx_sent_queue(dev_queue, skb->len,
doorbell = __netdev_tx_sent_queue(dev_queue, desc_cb->send_bytes,
netdev_xmit_more());
hns3_tx_doorbell(ring, ret, doorbell);
......@@ -2721,8 +2728,12 @@ static bool hns3_nic_reclaim_desc(struct hns3_enet_ring *ring,
break;
desc_cb = &ring->desc_cb[ntc];
(*pkts) += (desc_cb->type == DESC_TYPE_SKB);
(*bytes) += desc_cb->length;
if (desc_cb->type == DESC_TYPE_SKB) {
(*pkts)++;
(*bytes) += desc_cb->send_bytes;
}
/* desc_cb will be cleaned, after hnae3_free_buffer_detach */
hns3_free_buffer_detach(ring, ntc, budget);
......
......@@ -298,7 +298,12 @@ struct hns3_desc_cb {
/* priv data for the desc, e.g. skb when use with ip stack */
void *priv;
u32 page_offset;
union {
u32 page_offset; /* for rx */
u32 send_bytes; /* for tx */
};
u32 length; /* length of the buffer */
u16 reuse_flag;
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册