From: Herbert Xu <[EMAIL PROTECTED]>
Date: Mon, 22 Aug 2005 13:04:31 +1000
> > +
> > + fack_count += pcount;
> >
> > in_sack = !after(start_seq, TCP_SKB_CB(skb)->seq) &&
> > !before(end_seq, TCP_SKB_CB(skb)->end_seq);
>
> Maybe I'm missing something subtle here :) But it seems to me that
> your patch will end up only processing the second half of the result
> of tcp_fragment and skip the first bit.
Thinking about this some more... why would I need to use
anyting other than "SKB"? SKB is where the front part of
the split data is placed, and skb->next contains the
subsequent data.
So just using "skb" should be fine.
We also don't need to jump back to the top of the loop
because the first test:
/* The retransmission queue is always in order, so
* we can short-circuit the walk early.
*/
if (!before(TCP_SKB_CB(skb)->seq, end_seq))
break;
is by definition not going to pass.
The patch below should address the rest of your concerns.
A further simplification seems possible, in that all of the
tp->*_out modifications can be keyed upon the "skb_ever_sent"
arg to tcp_fragment().
diff --git a/include/net/tcp.h b/include/net/tcp.h
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -855,7 +855,7 @@ extern int tcp_retransmit_skb(struct soc
extern void tcp_xmit_retransmit_queue(struct sock *);
extern void tcp_simple_retransmit(struct sock *);
extern int tcp_trim_head(struct sock *, struct sk_buff *, u32);
-extern int tcp_fragment(struct sock *, struct sk_buff *, u32, unsigned int);
+extern int tcp_fragment(struct sock *, struct sk_buff *, u32, unsigned int,
int);
extern void tcp_send_probe0(struct sock *);
extern void tcp_send_partial(struct sock *);
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -978,7 +978,7 @@ tcp_sacktag_write_queue(struct sock *sk,
else
pkt_len = (end_seq -
TCP_SKB_CB(skb)->seq);
- if (tcp_fragment(sk, skb, pkt_len, pkt_len))
+ if (tcp_fragment(sk, skb, pkt_len, pkt_len, 1))
break;
pcount = tcp_skb_pcount(skb);
}
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -427,11 +427,11 @@ static void tcp_set_skb_tso_segs(struct
* packet to the list. This won't be called frequently, I hope.
* Remember, these are still headerless SKBs at this point.
*/
-int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int
mss_now)
+int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int
mss_now, int skb_ever_sent)
{
struct tcp_sock *tp = tcp_sk(sk);
struct sk_buff *buff;
- int nsize;
+ int nsize, old_factor;
u16 flags;
nsize = skb_headlen(skb) - len;
@@ -489,6 +489,8 @@ int tcp_fragment(struct sock *sk, struct
tp->left_out -= tcp_skb_pcount(skb);
}
+ old_factor = tcp_skb_pcount(skb);
+
/* Fix up tso_factor for both original and new SKB. */
tcp_set_skb_tso_segs(sk, skb, mss_now);
tcp_set_skb_tso_segs(sk, buff, mss_now);
@@ -503,6 +505,17 @@ int tcp_fragment(struct sock *sk, struct
tp->left_out += tcp_skb_pcount(buff);
}
+ if (skb_ever_sent) {
+ int diff = old_factor - tcp_skb_pcount(skb) -
+ tcp_skb_pcount(buff);
+ tp->packets_out -= diff;
+ if (diff > 0) {
+ tp->fackets_out -= diff;
+ if ((int)tp->fackets_out < 0)
+ tp->fackets_out = 0;
+ }
+ }
+
/* Link BUFF into the send queue. */
skb_header_release(buff);
__skb_append(skb, buff);
@@ -862,7 +875,7 @@ static int tso_fragment(struct sock *sk,
/* All of a TSO frame must be composed of paged data. */
if (skb->len != skb->data_len)
- return tcp_fragment(sk, skb, len, mss_now);
+ return tcp_fragment(sk, skb, len, mss_now, 0);
buff = sk_stream_alloc_pskb(sk, 0, 0, GFP_ATOMIC);
if (unlikely(buff == NULL))
@@ -1360,22 +1373,8 @@ int tcp_retransmit_skb(struct sock *sk,
return -EAGAIN;
if (skb->len > cur_mss) {
- int old_factor = tcp_skb_pcount(skb);
- int diff;
-
- if (tcp_fragment(sk, skb, cur_mss, cur_mss))
+ if (tcp_fragment(sk, skb, cur_mss, cur_mss, 1))
return -ENOMEM; /* We'll try again later. */
-
- /* New SKB created, account for it. */
- diff = old_factor - tcp_skb_pcount(skb) -
- tcp_skb_pcount(skb->next);
- tp->packets_out -= diff;
-
- if (diff > 0) {
- tp->fackets_out -= diff;
- if ((int)tp->fackets_out < 0)
- tp->fackets_out = 0;
- }
}
/* Collapse two adjacent packets if worthwhile and we can. */
@@ -1972,7 +1971,7 @@ int tcp_write_wakeup(struct sock *sk)
skb->len > mss) {
seg_size = min(seg_size, mss);
TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_PSH;
- if (tcp_fragment(sk, skb, seg_size, mss))
+ if (tcp_fragment(sk, skb, seg_size, mss, 0))
return -1;
} else if (!tcp_skb_pcount(skb))
tcp_set_skb_tso_segs(sk, skb, mss);
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
