#include <linux/module.h>
#include <linux/sysctl.h>
#include <linux/kernel.h>
+#include <linux/prefetch.h>
#include <net/dst.h>
#include <net/tcp.h>
#include <net/inet_common.h>
#include <linux/ipsec.h>
#include <asm/unaligned.h>
-#include <net/netdma.h>
#include <linux/errqueue.h>
int sysctl_tcp_timestamps __read_mostly = 1;
return packets_acked;
}
+static void tcp_ack_tstamp(struct sock *sk, struct sk_buff *skb,
+ u32 prior_snd_una)
+{
+ const struct skb_shared_info *shinfo;
+
+ /* Avoid cache line misses to get skb_shinfo() and shinfo->tx_flags */
+ if (likely(!(sk->sk_tsflags & SOF_TIMESTAMPING_TX_ACK)))
+ return;
+
+ shinfo = skb_shinfo(skb);
+ if ((shinfo->tx_flags & SKBTX_ACK_TSTAMP) &&
+ between(shinfo->tskey, prior_snd_una, tcp_sk(sk)->snd_una - 1))
+ __skb_tstamp_tx(skb, NULL, sk, SCM_TSTAMP_ACK);
+}
+
/* Remove acknowledged frames from the retransmission queue. If our packet
* is before the ack sequence we can discard it as it's confirmed to have
* arrived at the other end.
first_ackt.v64 = 0;
while ((skb = tcp_write_queue_head(sk)) && skb != tcp_send_head(sk)) {
- struct skb_shared_info *shinfo = skb_shinfo(skb);
struct tcp_skb_cb *scb = TCP_SKB_CB(skb);
u8 sacked = scb->sacked;
u32 acked_pcount;
- if (unlikely(shinfo->tx_flags & SKBTX_ACK_TSTAMP) &&
- between(shinfo->tskey, prior_snd_una, tp->snd_una - 1))
- __skb_tstamp_tx(skb, NULL, sk, SCM_TSTAMP_ACK);
+ tcp_ack_tstamp(sk, skb, prior_snd_una);
/* Determine how many packets and what bytes were acked, tso and else */
if (after(scb->end_seq, tp->snd_una)) {
fully_acked = false;
} else {
+ /* Speedup tcp_unlink_write_queue() and next loop */
+ prefetchw(skb->next);
acked_pcount = tcp_skb_pcount(skb);
}
- if (sacked & TCPCB_RETRANS) {
+ if (unlikely(sacked & TCPCB_RETRANS)) {
if (sacked & TCPCB_SACKED_RETRANS)
tp->retrans_out -= acked_pcount;
flag |= FLAG_RETRANS_DATA_ACKED;
* connection startup slow start one packet too
* quickly. This is severely frowned upon behavior.
*/
- if (!(scb->tcp_flags & TCPHDR_SYN)) {
+ if (likely(!(scb->tcp_flags & TCPHDR_SYN))) {
flag |= FLAG_DATA_ACKED;
} else {
flag |= FLAG_SYN_ACKED;
tcp_unlink_write_queue(skb, sk);
sk_wmem_free_skb(sk, skb);
- if (skb == tp->retransmit_skb_hint)
+ if (unlikely(skb == tp->retransmit_skb_hint))
tp->retransmit_skb_hint = NULL;
- if (skb == tp->lost_skb_hint)
+ if (unlikely(skb == tp->lost_skb_hint))
tp->lost_skb_hint = NULL;
}
flag |= FLAG_SACK_RENEGING;
skb_mstamp_get(&now);
- if (first_ackt.v64) {
+ if (likely(first_ackt.v64)) {
seq_rtt_us = skb_mstamp_us_delta(&now, &first_ackt);
ca_seq_rtt_us = skb_mstamp_us_delta(&now, &last_ackt);
}
int acked = 0; /* Number of packets newly acked */
long sack_rtt_us = -1L;
+ /* We very likely will need to access write queue head. */
+ prefetchw(sk->sk_write_queue.next);
+
/* If the ack is older than previous acks
* then we can probably ignore it.
*/
__tcp_checksum_complete_user(sk, skb);
}
-#ifdef CONFIG_NET_DMA
-static bool tcp_dma_try_early_copy(struct sock *sk, struct sk_buff *skb,
- int hlen)
-{
- struct tcp_sock *tp = tcp_sk(sk);
- int chunk = skb->len - hlen;
- int dma_cookie;
- bool copied_early = false;
-
- if (tp->ucopy.wakeup)
- return false;
-
- if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
- tp->ucopy.dma_chan = net_dma_find_channel();
-
- if (tp->ucopy.dma_chan && skb_csum_unnecessary(skb)) {
-
- dma_cookie = dma_skb_copy_datagram_iovec(tp->ucopy.dma_chan,
- skb, hlen,
- tp->ucopy.iov, chunk,
- tp->ucopy.pinned_list);
-
- if (dma_cookie < 0)
- goto out;
-
- tp->ucopy.dma_cookie = dma_cookie;
- copied_early = true;
-
- tp->ucopy.len -= chunk;
- tp->copied_seq += chunk;
- tcp_rcv_space_adjust(sk);
-
- if ((tp->ucopy.len == 0) ||
- (tcp_flag_word(tcp_hdr(skb)) & TCP_FLAG_PSH) ||
- (atomic_read(&sk->sk_rmem_alloc) > (sk->sk_rcvbuf >> 1))) {
- tp->ucopy.wakeup = 1;
- sk->sk_data_ready(sk);
- }
- } else if (chunk > 0) {
- tp->ucopy.wakeup = 1;
- sk->sk_data_ready(sk);
- }
-out:
- return copied_early;
-}
-#endif /* CONFIG_NET_DMA */
-
/* Does PAWS and seqno based validation of an incoming segment, flags will
* play significant role here.
*/
}
} else {
int eaten = 0;
- int copied_early = 0;
bool fragstolen = false;
- if (tp->copied_seq == tp->rcv_nxt &&
- len - tcp_header_len <= tp->ucopy.len) {
-#ifdef CONFIG_NET_DMA
- if (tp->ucopy.task == current &&
- sock_owned_by_user(sk) &&
- tcp_dma_try_early_copy(sk, skb, tcp_header_len)) {
- copied_early = 1;
- eaten = 1;
- }
-#endif
- if (tp->ucopy.task == current &&
- sock_owned_by_user(sk) && !copied_early) {
- __set_current_state(TASK_RUNNING);
+ if (tp->ucopy.task == current &&
+ tp->copied_seq == tp->rcv_nxt &&
+ len - tcp_header_len <= tp->ucopy.len &&
+ sock_owned_by_user(sk)) {
+ __set_current_state(TASK_RUNNING);
- if (!tcp_copy_to_iovec(sk, skb, tcp_header_len))
- eaten = 1;
- }
- if (eaten) {
+ if (!tcp_copy_to_iovec(sk, skb, tcp_header_len)) {
/* Predicted packet is in window by definition.
* seq == rcv_nxt and rcv_wup <= rcv_nxt.
* Hence, check seq<=rcv_wup reduces to:
__skb_pull(skb, tcp_header_len);
tp->rcv_nxt = TCP_SKB_CB(skb)->end_seq;
NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPHPHITSTOUSER);
+ eaten = 1;
}
- if (copied_early)
- tcp_cleanup_rbuf(sk, skb->len);
}
if (!eaten) {
if (tcp_checksum_complete_user(sk, skb))
goto no_ack;
}
- if (!copied_early || tp->rcv_nxt != tp->rcv_wup)
- __tcp_ack_snd_check(sk, 0);
+ __tcp_ack_snd_check(sk, 0);
no_ack:
-#ifdef CONFIG_NET_DMA
- if (copied_early)
- __skb_queue_tail(&sk->sk_async_wait_queue, skb);
- else
-#endif
if (eaten)
kfree_skb_partial(skb, fragstolen);
sk->sk_data_ready(sk);