1 files changed, 100 insertions, 27 deletions

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index e2b7b80..74998f2 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -263,7 +263,7 @@
 #include <net/tcp.h>
 #include <net/xfrm.h>
 #include <net/ip.h>
-
+#include <net/netdma.h>
 
 #include <asm/uaccess.h>
 #include <asm/ioctls.h>
@@ -622,14 +622,10 @@ ssize_t tcp_sendpage(struct socket *sock, struct page *page, int offset,
 	ssize_t res;
 	struct sock *sk = sock->sk;
 
-#define TCP_ZC_CSUM_FLAGS (NETIF_F_IP_CSUM | NETIF_F_NO_CSUM | NETIF_F_HW_CSUM)
-
 	if (!(sk->sk_route_caps & NETIF_F_SG) ||
-	    !(sk->sk_route_caps & TCP_ZC_CSUM_FLAGS))
+	    !(sk->sk_route_caps & NETIF_F_ALL_CSUM))
 		return sock_no_sendpage(sock, page, offset, size, flags);
 
-#undef TCP_ZC_CSUM_FLAGS
-
 	lock_sock(sk);
 	TCP_CHECK_TIMER(sk);
 	res = do_tcp_sendpages(sk, &page, offset, size, flags);
@@ -726,9 +722,7 @@ new_segment:
 				/*
 				 * Check whether we can use HW checksum.
 				 */
-				if (sk->sk_route_caps &
-				    (NETIF_F_IP_CSUM | NETIF_F_NO_CSUM |
-				     NETIF_F_HW_CSUM))
+				if (sk->sk_route_caps & NETIF_F_ALL_CSUM)
 					skb->ip_summed = CHECKSUM_HW;
 
 				skb_entail(sk, tp, skb);
@@ -937,7 +931,7 @@ static int tcp_recv_urg(struct sock *sk, long timeo,
  * calculation of whether or not we must ACK for the sake of
  * a window update.
  */
-static void cleanup_rbuf(struct sock *sk, int copied)
+void tcp_cleanup_rbuf(struct sock *sk, int copied)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
 	int time_to_ack = 0;
@@ -1072,11 +1066,11 @@ int tcp_read_sock(struct sock *sk, read_descriptor_t *desc,
 				break;
 		}
 		if (skb->h.th->fin) {
-			sk_eat_skb(sk, skb);
+			sk_eat_skb(sk, skb, 0);
 			++seq;
 			break;
 		}
-		sk_eat_skb(sk, skb);
+		sk_eat_skb(sk, skb, 0);
 		if (!desc->count)
 			break;
 	}
@@ -1086,7 +1080,7 @@ int tcp_read_sock(struct sock *sk, read_descriptor_t *desc,
 
 	/* Clean up data we have read: This will do ACK frames. */
 	if (copied)
-		cleanup_rbuf(sk, copied);
+		tcp_cleanup_rbuf(sk, copied);
 	return copied;
 }
 
@@ -1110,6 +1104,7 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 	int target;		/* Read at least this many bytes */
 	long timeo;
 	struct task_struct *user_recv = NULL;
+	int copied_early = 0;
 
 	lock_sock(sk);
 
@@ -1133,6 +1128,17 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 
 	target = sock_rcvlowat(sk, flags & MSG_WAITALL, len);
 
+#ifdef CONFIG_NET_DMA
+	tp->ucopy.dma_chan = NULL;
+	preempt_disable();
+	if ((len > sysctl_tcp_dma_copybreak) && !(flags & MSG_PEEK) &&
+	    !sysctl_tcp_low_latency && __get_cpu_var(softnet_data.net_dma)) {
+		preempt_enable_no_resched();
+		tp->ucopy.pinned_list = dma_pin_iovec_pages(msg->msg_iov, len);
+	} else
+		preempt_enable_no_resched();
+#endif
+
 	do {
 		struct sk_buff *skb;
 		u32 offset;
@@ -1220,7 +1226,7 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 			}
 		}
 
-		cleanup_rbuf(sk, copied);
+		tcp_cleanup_rbuf(sk, copied);
 
 		if (!sysctl_tcp_low_latency && tp->ucopy.task == user_recv) {
 			/* Install new reader */
@@ -1274,6 +1280,10 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 		} else
 			sk_wait_data(sk, &timeo);
 
+#ifdef CONFIG_NET_DMA
+		tp->ucopy.wakeup = 0;
+#endif
+
 		if (user_recv) {
 			int chunk;
 
@@ -1329,13 +1339,39 @@ do_prequeue:
 		}
 
 		if (!(flags & MSG_TRUNC)) {
-			err = skb_copy_datagram_iovec(skb, offset,
-						      msg->msg_iov, used);
-			if (err) {
-				/* Exception. Bailout! */
-				if (!copied)
-					copied = -EFAULT;
-				break;
+#ifdef CONFIG_NET_DMA
+			if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
+				tp->ucopy.dma_chan = get_softnet_dma();
+
+			if (tp->ucopy.dma_chan) {
+				tp->ucopy.dma_cookie = dma_skb_copy_datagram_iovec(
+					tp->ucopy.dma_chan, skb, offset,
+					msg->msg_iov, used,
+					tp->ucopy.pinned_list);
+
+				if (tp->ucopy.dma_cookie < 0) {
+
+					printk(KERN_ALERT "dma_cookie < 0\n");
+
+					/* Exception. Bailout! */
+					if (!copied)
+						copied = -EFAULT;
+					break;
+				}
+				if ((offset + used) == skb->len)
+					copied_early = 1;
+
+			} else
+#endif
+			{
+				err = skb_copy_datagram_iovec(skb, offset,
+						msg->msg_iov, used);
+				if (err) {
+					/* Exception. Bailout! */
+					if (!copied)
+						copied = -EFAULT;
+					break;
+				}
 			}
 		}
 
@@ -1355,15 +1391,19 @@ skip_copy:
 
 		if (skb->h.th->fin)
 			goto found_fin_ok;
-		if (!(flags & MSG_PEEK))
-			sk_eat_skb(sk, skb);
+		if (!(flags & MSG_PEEK)) {
+			sk_eat_skb(sk, skb, copied_early);
+			copied_early = 0;
+		}
 		continue;
 
 	found_fin_ok:
 		/* Process the FIN. */
 		++*seq;
-		if (!(flags & MSG_PEEK))
-			sk_eat_skb(sk, skb);
+		if (!(flags & MSG_PEEK)) {
+			sk_eat_skb(sk, skb, copied_early);
+			copied_early = 0;
+		}
 		break;
 	} while (len > 0);
 
@@ -1386,12 +1426,42 @@ skip_copy:
 		tp->ucopy.len = 0;
 	}
 
+#ifdef CONFIG_NET_DMA
+	if (tp->ucopy.dma_chan) {
+		struct sk_buff *skb;
+		dma_cookie_t done, used;
+
+		dma_async_memcpy_issue_pending(tp->ucopy.dma_chan);
+
+		while (dma_async_memcpy_complete(tp->ucopy.dma_chan,
+		                                 tp->ucopy.dma_cookie, &done,
+		                                 &used) == DMA_IN_PROGRESS) {
+			/* do partial cleanup of sk_async_wait_queue */
+			while ((skb = skb_peek(&sk->sk_async_wait_queue)) &&
+			       (dma_async_is_complete(skb->dma_cookie, done,
+			                              used) == DMA_SUCCESS)) {
+				__skb_dequeue(&sk->sk_async_wait_queue);
+				kfree_skb(skb);
+			}
+		}
+
+		/* Safe to free early-copied skbs now */
+		__skb_queue_purge(&sk->sk_async_wait_queue);
+		dma_chan_put(tp->ucopy.dma_chan);
+		tp->ucopy.dma_chan = NULL;
+	}
+	if (tp->ucopy.pinned_list) {
+		dma_unpin_iovec_pages(tp->ucopy.pinned_list);
+		tp->ucopy.pinned_list = NULL;
+	}
+#endif
+
 	/* According to UNIX98, msg_name/msg_namelen are ignored
 	 * on connected socket. I was just happy when found this 8) --ANK
 	 */
 
 	/* Clean up data we have read: This will do ACK frames. */
-	cleanup_rbuf(sk, copied);
+	tcp_cleanup_rbuf(sk, copied);
 
 	TCP_CHECK_TIMER(sk);
 	release_sock(sk);
@@ -1658,6 +1728,9 @@ int tcp_disconnect(struct sock *sk, int flags)
 	__skb_queue_purge(&sk->sk_receive_queue);
 	sk_stream_writequeue_purge(sk);
 	__skb_queue_purge(&tp->out_of_order_queue);
+#ifdef CONFIG_NET_DMA
+	__skb_queue_purge(&sk->sk_async_wait_queue);
+#endif
 
 	inet->dport = 0;
 
@@ -1858,7 +1931,7 @@ static int do_tcp_setsockopt(struct sock *sk, int level,
 			    (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT) &&
 			    inet_csk_ack_scheduled(sk)) {
 				icsk->icsk_ack.pending |= ICSK_ACK_PUSHED;
-				cleanup_rbuf(sk, 1);
+				tcp_cleanup_rbuf(sk, 1);
 				if (!(val & 1))
 					icsk->icsk_ack.pingpong = 1;
 			}