diff options
| author | Eric Dumazet <edumazet@google.com> | 2012-07-31 05:45:30 +0000 |
|---|---|---|
| committer | David S. Miller <davem@davemloft.net> | 2012-07-31 14:41:39 -0700 |
| commit | d26b3a7c4b3b26319f18bb645de93eba8f4bdcd5 (patch) | |
| tree | ca86c03450fafdc89dac98ce403b1906fcaa025d /include/net/ip_fib.h | |
| parent | 54764bb647b2e847c512acf8d443df965da35000 (diff) | |
| download | op-kernel-dev-d26b3a7c4b3b26319f18bb645de93eba8f4bdcd5.zip op-kernel-dev-d26b3a7c4b3b26319f18bb645de93eba8f4bdcd5.tar.gz | |
ipv4: percpu nh_rth_output cache
Input path is mostly run under RCU and doesnt touch dst refcnt
But output path on forwarding or UDP workloads hits
badly dst refcount, and we have lot of false sharing, for example
in ipv4_mtu() when reading rt->rt_pmtu
Using a percpu cache for nh_rth_output gives a nice performance
increase at a small cost.
24 udpflood test on my 24 cpu machine (dummy0 output device)
(each process sends 1.000.000 udp frames, 24 processes are started)
before : 5.24 s
after : 2.06 s
For reference, time on linux-3.5 : 6.60 s
Signed-off-by: Eric Dumazet <edumazet@google.com>
Tested-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'include/net/ip_fib.h')
| -rw-r--r-- | include/net/ip_fib.h | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h index e521a03..e331746 100644 --- a/include/net/ip_fib.h +++ b/include/net/ip_fib.h @@ -21,6 +21,7 @@ #include <linux/rcupdate.h> #include <net/fib_rules.h> #include <net/inetpeer.h> +#include <linux/percpu.h> struct fib_config { u8 fc_dst_len; @@ -81,7 +82,7 @@ struct fib_nh { __be32 nh_gw; __be32 nh_saddr; int nh_saddr_genid; - struct rtable __rcu *nh_rth_output; + struct rtable __rcu * __percpu *nh_pcpu_rth_output; struct rtable __rcu *nh_rth_input; struct fnhe_hash_bucket *nh_exceptions; }; |
