From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: [PATCH net-next-2.6] inetpeer: do not use zero refcnt for freed entries Date: Wed, 16 Jun 2010 04:45:24 +0200 Message-ID: <1276656324.19249.39.camel@edumazet-laptop> References: <1276626194.2541.186.camel@edumazet-laptop> <20100615.142506.02275206.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: netdev@vger.kernel.org, paulmck@linux.vnet.ibm.com To: David Miller Return-path: Received: from mail-ww0-f46.google.com ([74.125.82.46]:59433 "EHLO mail-ww0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755151Ab0FPCpa (ORCPT ); Tue, 15 Jun 2010 22:45:30 -0400 Received: by wwb18 with SMTP id 18so5297487wwb.19 for ; Tue, 15 Jun 2010 19:45:28 -0700 (PDT) In-Reply-To: <20100615.142506.02275206.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: Le mardi 15 juin 2010 =C3=A0 14:25 -0700, David Miller a =C3=A9crit : > From: Eric Dumazet > Date: Tue, 15 Jun 2010 20:23:14 +0200 >=20 > > inetpeer currently uses an AVL tree protected by an rwlock. > >=20 > > It's possible to make most lookups use RCU > ... > > Signed-off-by: Eric Dumazet >=20 > Applied, nice work Eric. Thanks David ! Re-reading patch I realize refcnt is expected to be 0 for unused entrie= s (obviously), so we should use a different marker for 'about to be freed= ' ones. Thanks [PATCH net-next-2.6] inetpeer: do not use zero refcnt for freed entries =46ollowup of commit aa1039e73cc2 (inetpeer: RCU conversion) Unused inet_peer entries have a null refcnt. Using atomic_inc_not_zero() in rcu lookups is not going to work for them, and slow path is taken. =46ix this using -1 marker instead of 0 for deleted entries. Signed-off-by: Eric Dumazet --- net/ipv4/inetpeer.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/net/ipv4/inetpeer.c b/net/ipv4/inetpeer.c index 58fbc7e..39a14ba 100644 --- a/net/ipv4/inetpeer.c +++ b/net/ipv4/inetpeer.c @@ -187,7 +187,12 @@ static struct inet_peer *lookup_rcu_bh(__be32 dadd= r) =20 while (u !=3D peer_avl_empty) { if (daddr =3D=3D u->v4daddr) { - if (unlikely(!atomic_inc_not_zero(&u->refcnt))) + /* Before taking a reference, check if this entry was + * deleted, unlink_from_pool() sets refcnt=3D-1 to make + * distinction between an unused entry (refcnt=3D0) and + * a freed one. + */ + if (unlikely(!atomic_add_unless(&u->refcnt, 1, -1))) u =3D NULL; return u; } @@ -322,8 +327,9 @@ static void unlink_from_pool(struct inet_peer *p) * in cleanup() function to prevent sudden disappearing. If we can * atomically (because of lockless readers) take this last reference, * it's safe to remove the node and free it later. + * We use refcnt=3D-1 to alert lockless readers this entry is deleted= =2E */ - if (atomic_cmpxchg(&p->refcnt, 1, 0) =3D=3D 1) { + if (atomic_cmpxchg(&p->refcnt, 1, -1) =3D=3D 1) { struct inet_peer **stack[PEER_MAXDEPTH]; struct inet_peer ***stackptr, ***delp; if (lookup(p->v4daddr, stack) !=3D p)