]> Pileus Git - ~andy/linux/commitdiff
ipv4: gso: send_check() & segment() cleanups
authorEric Dumazet <edumazet@google.com>
Fri, 18 Oct 2013 20:13:27 +0000 (13:13 -0700)
committerDavid S. Miller <davem@davemloft.net>
Sat, 19 Oct 2013 23:11:56 +0000 (19:11 -0400)
inet_gso_segment() and inet_gso_send_check() are called by
skb_mac_gso_segment() under rcu lock, no need to use
rcu_read_lock() / rcu_read_unlock()

Avoid calling ip_hdr() twice per function.

We can use ip_send_check() helper.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/af_inet.c

index 35913fb77dc8672ac30518078eaa08378d32617d..4f8cd4fc451dba1d0358d516426226fb0b8e2600 100644 (file)
@@ -1254,20 +1254,19 @@ static int inet_gso_send_check(struct sk_buff *skb)
        if (ihl < sizeof(*iph))
                goto out;
 
+       proto = iph->protocol;
+
+       /* Warning: after this point, iph might be no longer valid */
        if (unlikely(!pskb_may_pull(skb, ihl)))
                goto out;
-
        __skb_pull(skb, ihl);
+
        skb_reset_transport_header(skb);
-       iph = ip_hdr(skb);
-       proto = iph->protocol;
        err = -EPROTONOSUPPORT;
 
-       rcu_read_lock();
        ops = rcu_dereference(inet_offloads[proto]);
        if (likely(ops && ops->callbacks.gso_send_check))
                err = ops->callbacks.gso_send_check(skb);
-       rcu_read_unlock();
 
 out:
        return err;
@@ -1305,23 +1304,23 @@ static struct sk_buff *inet_gso_segment(struct sk_buff *skb,
        if (ihl < sizeof(*iph))
                goto out;
 
+       id = ntohs(iph->id);
+       proto = iph->protocol;
+
+       /* Warning: after this point, iph might be no longer valid */
        if (unlikely(!pskb_may_pull(skb, ihl)))
                goto out;
+       __skb_pull(skb, ihl);
 
        tunnel = !!skb->encapsulation;
 
-       __skb_pull(skb, ihl);
        skb_reset_transport_header(skb);
-       iph = ip_hdr(skb);
-       id = ntohs(iph->id);
-       proto = iph->protocol;
+
        segs = ERR_PTR(-EPROTONOSUPPORT);
 
-       rcu_read_lock();
        ops = rcu_dereference(inet_offloads[proto]);
        if (likely(ops && ops->callbacks.gso_segment))
                segs = ops->callbacks.gso_segment(skb, features);
-       rcu_read_unlock();
 
        if (IS_ERR_OR_NULL(segs))
                goto out;
@@ -1339,8 +1338,7 @@ static struct sk_buff *inet_gso_segment(struct sk_buff *skb,
                        iph->id = htons(id++);
                }
                iph->tot_len = htons(skb->len - skb->mac_len);
-               iph->check = 0;
-               iph->check = ip_fast_csum(skb_network_header(skb), iph->ihl);
+               ip_send_check(iph);
        } while ((skb = skb->next));
 
 out: