提交 d84130a1 编写于 作者: H hayeswang 提交者: David S. Miller

r8152: reduce the frequency of spin_lock

Replace getting one item from a list with getting the whole list one time.
Signed-off-by: NHayes Wang <hayeswang@realtek.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
上级 8a91c824
......@@ -1299,9 +1299,16 @@ r8152_tx_csum(struct r8152 *tp, struct tx_desc *desc, struct sk_buff *skb)
static int r8152_tx_agg_fill(struct r8152 *tp, struct tx_agg *agg)
{
struct sk_buff_head skb_head, *tx_queue = &tp->tx_queue;
unsigned long flags;
int remain;
u8 *tx_data;
__skb_queue_head_init(&skb_head);
spin_lock_irqsave(&tx_queue->lock, flags);
skb_queue_splice_init(tx_queue, &skb_head);
spin_unlock_irqrestore(&tx_queue->lock, flags);
tx_data = agg->head;
agg->skb_num = agg->skb_len = 0;
remain = rx_buf_sz;
......@@ -1311,14 +1318,14 @@ static int r8152_tx_agg_fill(struct r8152 *tp, struct tx_agg *agg)
struct sk_buff *skb;
unsigned int len;
skb = skb_dequeue(&tp->tx_queue);
skb = __skb_dequeue(&skb_head);
if (!skb)
break;
remain -= sizeof(*tx_desc);
len = skb->len;
if (remain < len) {
skb_queue_head(&tp->tx_queue, skb);
__skb_queue_head(&skb_head, skb);
break;
}
......@@ -1336,6 +1343,12 @@ static int r8152_tx_agg_fill(struct r8152 *tp, struct tx_agg *agg)
remain = rx_buf_sz - (int)(tx_agg_align(tx_data) - agg->head);
}
if (!skb_queue_empty(&skb_head)) {
spin_lock_irqsave(&tx_queue->lock, flags);
skb_queue_splice(&skb_head, tx_queue);
spin_unlock_irqrestore(&tx_queue->lock, flags);
}
netif_tx_lock(tp->netdev);
if (netif_queue_stopped(tp->netdev) &&
......@@ -1354,10 +1367,17 @@ static int r8152_tx_agg_fill(struct r8152 *tp, struct tx_agg *agg)
static void rx_bottom(struct r8152 *tp)
{
unsigned long flags;
struct list_head *cursor, *next;
struct list_head *cursor, *next, rx_queue;
if (list_empty(&tp->rx_done))
return;
INIT_LIST_HEAD(&rx_queue);
spin_lock_irqsave(&tp->rx_lock, flags);
list_for_each_safe(cursor, next, &tp->rx_done) {
list_splice_init(&tp->rx_done, &rx_queue);
spin_unlock_irqrestore(&tp->rx_lock, flags);
list_for_each_safe(cursor, next, &rx_queue) {
struct rx_desc *rx_desc;
struct rx_agg *agg;
int len_used = 0;
......@@ -1366,7 +1386,6 @@ static void rx_bottom(struct r8152 *tp)
int ret;
list_del_init(cursor);
spin_unlock_irqrestore(&tp->rx_lock, flags);
agg = list_entry(cursor, struct rx_agg, list);
urb = agg->urb;
......@@ -1416,13 +1435,13 @@ static void rx_bottom(struct r8152 *tp)
submit:
ret = r8152_submit_rx(tp, agg, GFP_ATOMIC);
spin_lock_irqsave(&tp->rx_lock, flags);
if (ret && ret != -ENODEV) {
list_add_tail(&agg->list, next);
spin_lock_irqsave(&tp->rx_lock, flags);
list_add_tail(&agg->list, &tp->rx_done);
spin_unlock_irqrestore(&tp->rx_lock, flags);
tasklet_schedule(&tp->tl);
}
}
spin_unlock_irqrestore(&tp->rx_lock, flags);
}
static void tx_bottom(struct r8152 *tp)
......@@ -1496,9 +1515,19 @@ int r8152_submit_rx(struct r8152 *tp, struct rx_agg *agg, gfp_t mem_flags)
static void rtl_drop_queued_tx(struct r8152 *tp)
{
struct net_device_stats *stats = &tp->netdev->stats;
struct sk_buff_head skb_head, *tx_queue = &tp->tx_queue;
unsigned long flags;
struct sk_buff *skb;
while ((skb = skb_dequeue(&tp->tx_queue))) {
if (skb_queue_empty(tx_queue))
return;
__skb_queue_head_init(&skb_head);
spin_lock_irqsave(&tx_queue->lock, flags);
skb_queue_splice_init(tx_queue, &skb_head);
spin_unlock_irqrestore(&tx_queue->lock, flags);
while ((skb = __skb_dequeue(&skb_head))) {
dev_kfree_skb(skb);
stats->tx_dropped++;
}
......
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