OpenE2K
/
linux
6
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

net: sock_rps_record_flow() is for connected sockets 

Paolo noticed a cache line miss in UDP recvmsg() to access
sk_rxhash, sharing a cache line with sk_drops.

sk_drops might be heavily incremented by cpus handling a flood targeting
this socket.

We might place sk_drops on a separate cache line, but lets try
to avoid wasting 64 bytes per socket just for this, since we have
other bottlenecks to take care of.

sock_rps_record_flow() should only access sk_rxhash for connected
flows.

Testing sk_state for TCP_ESTABLISHED covers most of the cases for
connected sockets, for a zero cost, since system calls using
sock_rps_record_flow() also access sk->sk_prot which is on the
same cache line.

A follow up patch will provide a static_key (Jump Label) since most
hosts do not even use RFS.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Eric Dumazet 2016-12-06 19:32:50 -08:00 committed by David S. Miller
parent d4aea20d88
commit 5b8e2f61b9
1 changed files with 11 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
include/net/sock.h
View File

@ -913,7 +913,17 @@ static inline void sock_rps_record_flow_hash(__u32 hash)
static inline void sock_rps_record_flow(const struct sock *sk)
{
#ifdef CONFIG_RPS
sock_rps_record_flow_hash(sk->sk_rxhash);
/* Reading sk->sk_rxhash might incur an expensive cache line miss.
*
* TCP_ESTABLISHED does cover almost all states where RFS
* might be useful, and is cheaper [1] than testing :
* IPv4: inet_sk(sk)->inet_daddr
* IPv6: ipv6_addr_any(&sk->sk_v6_daddr)
* OR an additional socket flag
* [1] : sk_state and sk_prot are in the same cache line.
*/
if (sk->sk_state == TCP_ESTABLISHED)
sock_rps_record_flow_hash(sk->sk_rxhash);
#endif
}