diff options
author | Eric Dumazet <dada1@cosmosbay.com> | 2007-11-07 05:40:20 -0500 |
---|---|---|
committer | David S. Miller <davem@sunset.davemloft.net> | 2007-11-07 07:15:11 -0500 |
commit | 230140cffa7feae90ad50bf259db1fa07674f3a7 (patch) | |
tree | 815472add31606423a508a17806b7884f0ab3e2e /net/ipv4/inet_hashtables.c | |
parent | efac52762b1e3fe3035d29e82d8ee1aebc45e4a7 (diff) |
[INET]: Remove per bucket rwlock in tcp/dccp ehash table.
As done two years ago on IP route cache table (commit
22c047ccbc68fa8f3fa57f0e8f906479a062c426) , we can avoid using one
lock per hash bucket for the huge TCP/DCCP hash tables.
On a typical x86_64 platform, this saves about 2MB or 4MB of ram, for
litle performance differences. (we hit a different cache line for the
rwlock, but then the bucket cache line have a better sharing factor
among cpus, since we dirty it less often). For netstat or ss commands
that want a full scan of hash table, we perform fewer memory accesses.
Using a 'small' table of hashed rwlocks should be more than enough to
provide correct SMP concurrency between different buckets, without
using too much memory. Sizing of this table depends on
num_possible_cpus() and various CONFIG settings.
This patch provides some locking abstraction that may ease a future
work using a different model for TCP/DCCP table.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Acked-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/ipv4/inet_hashtables.c')
-rw-r--r-- | net/ipv4/inet_hashtables.c | 7 |
1 files changed, 4 insertions, 3 deletions
diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index 16eecc7046a3..67704da04fc4 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c | |||
@@ -204,12 +204,13 @@ static int __inet_check_established(struct inet_timewait_death_row *death_row, | |||
204 | const __portpair ports = INET_COMBINED_PORTS(inet->dport, lport); | 204 | const __portpair ports = INET_COMBINED_PORTS(inet->dport, lport); |
205 | unsigned int hash = inet_ehashfn(daddr, lport, saddr, inet->dport); | 205 | unsigned int hash = inet_ehashfn(daddr, lport, saddr, inet->dport); |
206 | struct inet_ehash_bucket *head = inet_ehash_bucket(hinfo, hash); | 206 | struct inet_ehash_bucket *head = inet_ehash_bucket(hinfo, hash); |
207 | rwlock_t *lock = inet_ehash_lockp(hinfo, hash); | ||
207 | struct sock *sk2; | 208 | struct sock *sk2; |
208 | const struct hlist_node *node; | 209 | const struct hlist_node *node; |
209 | struct inet_timewait_sock *tw; | 210 | struct inet_timewait_sock *tw; |
210 | 211 | ||
211 | prefetch(head->chain.first); | 212 | prefetch(head->chain.first); |
212 | write_lock(&head->lock); | 213 | write_lock(lock); |
213 | 214 | ||
214 | /* Check TIME-WAIT sockets first. */ | 215 | /* Check TIME-WAIT sockets first. */ |
215 | sk_for_each(sk2, node, &head->twchain) { | 216 | sk_for_each(sk2, node, &head->twchain) { |
@@ -239,7 +240,7 @@ unique: | |||
239 | BUG_TRAP(sk_unhashed(sk)); | 240 | BUG_TRAP(sk_unhashed(sk)); |
240 | __sk_add_node(sk, &head->chain); | 241 | __sk_add_node(sk, &head->chain); |
241 | sock_prot_inc_use(sk->sk_prot); | 242 | sock_prot_inc_use(sk->sk_prot); |
242 | write_unlock(&head->lock); | 243 | write_unlock(lock); |
243 | 244 | ||
244 | if (twp) { | 245 | if (twp) { |
245 | *twp = tw; | 246 | *twp = tw; |
@@ -255,7 +256,7 @@ unique: | |||
255 | return 0; | 256 | return 0; |
256 | 257 | ||
257 | not_unique: | 258 | not_unique: |
258 | write_unlock(&head->lock); | 259 | write_unlock(lock); |
259 | return -EADDRNOTAVAIL; | 260 | return -EADDRNOTAVAIL; |
260 | } | 261 | } |
261 | 262 | ||