diff options
author | Eric Dumazet <dada1@cosmosbay.com> | 2008-11-24 18:52:46 -0500 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2008-11-24 18:52:46 -0500 |
commit | 2e77d89b2fa8e3f8325b8ce7893ec3645f41aff5 (patch) | |
tree | ae40aa75449f705bd166630f9bcb5f41373d8248 /include/net/ip.h | |
parent | 4db0acf3c0afbbbb2ae35a65f8896ca6655a47ec (diff) |
net: avoid a pair of dst_hold()/dst_release() in ip_append_data()
We can reduce pressure on dst entry refcount that slowdown UDP transmit
path on SMP machines. This pressure is visible on RTP servers when
delivering content to mediagateways, especially big ones, handling
thousand of streams. Several cpus send UDP frames to the same
destination, hence use the same dst entry.
This patch makes ip_append_data() eventually steal the refcount its
callers had to take on the dst entry.
This doesnt avoid all refcounting, but still gives speedups on SMP,
on UDP/RAW transmit path
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'include/net/ip.h')
-rw-r--r-- | include/net/ip.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/include/net/ip.h b/include/net/ip.h index bc026ecb513f..ddef10c22e3a 100644 --- a/include/net/ip.h +++ b/include/net/ip.h | |||
@@ -110,7 +110,7 @@ extern int ip_append_data(struct sock *sk, | |||
110 | int odd, struct sk_buff *skb), | 110 | int odd, struct sk_buff *skb), |
111 | void *from, int len, int protolen, | 111 | void *from, int len, int protolen, |
112 | struct ipcm_cookie *ipc, | 112 | struct ipcm_cookie *ipc, |
113 | struct rtable *rt, | 113 | struct rtable **rt, |
114 | unsigned int flags); | 114 | unsigned int flags); |
115 | extern int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb); | 115 | extern int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb); |
116 | extern ssize_t ip_append_page(struct sock *sk, struct page *page, | 116 | extern ssize_t ip_append_page(struct sock *sk, struct page *page, |