diff options
| author | Eric Dumazet <edumazet@google.com> | 2012-06-08 05:03:21 +0000 |
|---|---|---|
| committer | David S. Miller <davem@davemloft.net> | 2012-06-08 14:27:23 -0700 |
| commit | 7123aaa3a1416529ce461e98108e6b343b294643 (patch) | |
| tree | 094070a521666979523411422109cce72a1db8c4 /include/net/af_unix.h | |
| parent | 54db0cc2ba0d38166acc2d6bae21721405305537 (diff) | |
| download | op-kernel-dev-7123aaa3a1416529ce461e98108e6b343b294643.zip op-kernel-dev-7123aaa3a1416529ce461e98108e6b343b294643.tar.gz | |
af_unix: speedup /proc/net/unix
/proc/net/unix has quadratic behavior, and can hold unix_table_lock for
a while if high number of unix sockets are alive. (90 ms for 200k
sockets...)
We already have a hash table, so its quite easy to use it.
Problem is unbound sockets are still hashed in a single hash slot
(unix_socket_table[UNIX_HASH_TABLE])
This patch also spreads unbound sockets to 256 hash slots, to speedup
both /proc/net/unix and unix_diag.
Time to read /proc/net/unix with 200k unix sockets :
(time dd if=/proc/net/unix of=/dev/null bs=4k)
before : 520 secs
after : 2 secs
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'include/net/af_unix.h')
| -rw-r--r-- | include/net/af_unix.h | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/include/net/af_unix.h b/include/net/af_unix.h index 2ee33da..b5f8988 100644 --- a/include/net/af_unix.h +++ b/include/net/af_unix.h @@ -14,10 +14,11 @@ extern struct sock *unix_get_socket(struct file *filp); extern struct sock *unix_peer_get(struct sock *); #define UNIX_HASH_SIZE 256 +#define UNIX_HASH_BITS 8 extern unsigned int unix_tot_inflight; extern spinlock_t unix_table_lock; -extern struct hlist_head unix_socket_table[UNIX_HASH_SIZE + 1]; +extern struct hlist_head unix_socket_table[2 * UNIX_HASH_SIZE]; struct unix_address { atomic_t refcnt; |
