From patchwork Sat Oct 27 16:05:39 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jozsef Kadlecsik X-Patchwork-Id: 989916 X-Patchwork-Delegate: pablo@netfilter.org Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=none (mailfrom) smtp.mailfrom=vger.kernel.org (client-ip=209.132.180.67; helo=vger.kernel.org; envelope-from=netfilter-devel-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dmarc=pass (p=none dis=none) header.from=blackhole.kfki.hu Authentication-Results: ozlabs.org; dkim=pass (1024-bit key; unprotected) header.d=blackhole.kfki.hu header.i=@blackhole.kfki.hu header.b="MkRpumrg"; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id 42j5Ls0rpjz9sCc for ; Sun, 28 Oct 2018 03:05:57 +1100 (AEDT) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1728765AbeJ1ArP (ORCPT ); Sat, 27 Oct 2018 20:47:15 -0400 Received: from smtp-out.kfki.hu ([148.6.0.45]:33273 "EHLO smtp-out.kfki.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1728739AbeJ1ArP (ORCPT ); Sat, 27 Oct 2018 20:47:15 -0400 Received: from localhost (localhost [127.0.0.1]) by smtp0.kfki.hu (Postfix) with ESMTP id 1D011674010E; Sat, 27 Oct 2018 18:05:46 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d= blackhole.kfki.hu; h=references:in-reply-to:x-mailer:message-id :date:date:from:from:received:received:received; s=20151130; t= 1540656344; x=1542470745; bh=EzxHAErf6E1G87+4yUJdvLaBbFi/2d3jdh/ ZDS+EUtA=; b=MkRpumrg9QYEfbUMFcQ83797vV5LPnncDQdvjYnlcL8GGjAve1Z ppb+i3OdH5tDHd6TC6AIrd3DoBznaCoEuqOnnQYOIViOZOu6WANPo75wRImBtmS8 BT8pWZj/lzavTsKIlfPVbgZOsohTE1r4k633H8c2ot9m0GdN4oavZ470= X-Virus-Scanned: Debian amavisd-new at smtp0.kfki.hu Received: from smtp0.kfki.hu ([127.0.0.1]) by localhost (smtp0.kfki.hu [127.0.0.1]) (amavisd-new, port 10026) with ESMTP; Sat, 27 Oct 2018 18:05:44 +0200 (CEST) Received: from blackhole.kfki.hu (blackhole.kfki.hu [IPv6:2001:738:5001:1::240:2]) by smtp0.kfki.hu (Postfix) with ESMTP id EB593674010D; Sat, 27 Oct 2018 18:05:43 +0200 (CEST) Received: by blackhole.kfki.hu (Postfix, from userid 1000) id C93DE20576; Sat, 27 Oct 2018 18:05:43 +0200 (CEST) From: Jozsef Kadlecsik To: netfilter-devel@vger.kernel.org Cc: Pablo Neira Ayuso Subject: [PATCH 1/5] netfilter: ipset: list:set: Decrease refcount synchronously on deletion and replace Date: Sat, 27 Oct 2018 18:05:39 +0200 Message-Id: <1540656343-822-2-git-send-email-kadlec@blackhole.kfki.hu> X-Mailer: git-send-email 2.1.4 In-Reply-To: <1540656343-822-1-git-send-email-kadlec@blackhole.kfki.hu> References: <1540656343-822-1-git-send-email-kadlec@blackhole.kfki.hu> Sender: netfilter-devel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netfilter-devel@vger.kernel.org From: Stefano Brivio Commit 45040978c899 ("netfilter: ipset: Fix set:list type crash when flush/dump set in parallel") postponed decreasing set reference counters to the RCU callback. An 'ipset del' command can terminate before the RCU grace period is elapsed, and if sets are listed before then, the reference counter shown in userspace will be wrong: # ipset create h hash:ip; ipset create l list:set; ipset add l # ipset del l h; ipset list h Name: h Type: hash:ip Revision: 4 Header: family inet hashsize 1024 maxelem 65536 Size in memory: 88 References: 1 Number of entries: 0 Members: # sleep 1; ipset list h Name: h Type: hash:ip Revision: 4 Header: family inet hashsize 1024 maxelem 65536 Size in memory: 88 References: 0 Number of entries: 0 Members: Fix this by making the reference count update synchronous again. As a result, when sets are listed, ip_set_name_byindex() might now fetch a set whose reference count is already zero. Instead of relying on the reference count to protect against concurrent set renaming, grab ip_set_ref_lock as reader and copy the name, while holding the same lock in ip_set_rename() as writer instead. Reported-by: Li Shuang Fixes: 45040978c899 ("netfilter: ipset: Fix set:list type crash when flush/dump set in parallel") Signed-off-by: Stefano Brivio Signed-off-by: Jozsef Kadlecsik --- include/linux/netfilter/ipset/ip_set.h | 2 +- net/netfilter/ipset/ip_set_core.c | 23 +++++++++++------------ net/netfilter/ipset/ip_set_list_set.c | 17 +++++++++++------ 3 files changed, 23 insertions(+), 19 deletions(-) diff --git a/include/linux/netfilter/ipset/ip_set.h b/include/linux/netfilter/ipset/ip_set.h index 34fc80f..1d100ef 100644 --- a/include/linux/netfilter/ipset/ip_set.h +++ b/include/linux/netfilter/ipset/ip_set.h @@ -314,7 +314,7 @@ enum { extern ip_set_id_t ip_set_get_byname(struct net *net, const char *name, struct ip_set **set); extern void ip_set_put_byindex(struct net *net, ip_set_id_t index); -extern const char *ip_set_name_byindex(struct net *net, ip_set_id_t index); +extern void ip_set_name_byindex(struct net *net, ip_set_id_t index, char *name); extern ip_set_id_t ip_set_nfnl_get_byindex(struct net *net, ip_set_id_t index); extern void ip_set_nfnl_put(struct net *net, ip_set_id_t index); diff --git a/net/netfilter/ipset/ip_set_core.c b/net/netfilter/ipset/ip_set_core.c index bc4bd247..fa15a83 100644 --- a/net/netfilter/ipset/ip_set_core.c +++ b/net/netfilter/ipset/ip_set_core.c @@ -693,21 +693,20 @@ ip_set_put_byindex(struct net *net, ip_set_id_t index) EXPORT_SYMBOL_GPL(ip_set_put_byindex); /* Get the name of a set behind a set index. - * We assume the set is referenced, so it does exist and - * can't be destroyed. The set cannot be renamed due to - * the referencing either. - * + * Set itself is protected by RCU, but its name isn't: to protect against + * renaming, grab ip_set_ref_lock as reader (see ip_set_rename()) and copy the + * name. */ -const char * -ip_set_name_byindex(struct net *net, ip_set_id_t index) +void +ip_set_name_byindex(struct net *net, ip_set_id_t index, char *name) { - const struct ip_set *set = ip_set_rcu_get(net, index); + struct ip_set *set = ip_set_rcu_get(net, index); BUG_ON(!set); - BUG_ON(set->ref == 0); - /* Referenced, so it's safe */ - return set->name; + read_lock_bh(&ip_set_ref_lock); + strncpy(name, set->name, IPSET_MAXNAMELEN); + read_unlock_bh(&ip_set_ref_lock); } EXPORT_SYMBOL_GPL(ip_set_name_byindex); @@ -1153,7 +1152,7 @@ static int ip_set_rename(struct net *net, struct sock *ctnl, if (!set) return -ENOENT; - read_lock_bh(&ip_set_ref_lock); + write_lock_bh(&ip_set_ref_lock); if (set->ref != 0) { ret = -IPSET_ERR_REFERENCED; goto out; @@ -1170,7 +1169,7 @@ static int ip_set_rename(struct net *net, struct sock *ctnl, strncpy(set->name, name2, IPSET_MAXNAMELEN); out: - read_unlock_bh(&ip_set_ref_lock); + write_unlock_bh(&ip_set_ref_lock); return ret; } diff --git a/net/netfilter/ipset/ip_set_list_set.c b/net/netfilter/ipset/ip_set_list_set.c index 072a658..4eef55d 100644 --- a/net/netfilter/ipset/ip_set_list_set.c +++ b/net/netfilter/ipset/ip_set_list_set.c @@ -148,9 +148,7 @@ __list_set_del_rcu(struct rcu_head * rcu) { struct set_elem *e = container_of(rcu, struct set_elem, rcu); struct ip_set *set = e->set; - struct list_set *map = set->data; - ip_set_put_byindex(map->net, e->id); ip_set_ext_destroy(set, e); kfree(e); } @@ -158,15 +156,21 @@ __list_set_del_rcu(struct rcu_head * rcu) static inline void list_set_del(struct ip_set *set, struct set_elem *e) { + struct list_set *map = set->data; + set->elements--; list_del_rcu(&e->list); + ip_set_put_byindex(map->net, e->id); call_rcu(&e->rcu, __list_set_del_rcu); } static inline void -list_set_replace(struct set_elem *e, struct set_elem *old) +list_set_replace(struct ip_set *set, struct set_elem *e, struct set_elem *old) { + struct list_set *map = set->data; + list_replace_rcu(&old->list, &e->list); + ip_set_put_byindex(map->net, old->id); call_rcu(&old->rcu, __list_set_del_rcu); } @@ -298,7 +302,7 @@ list_set_uadd(struct ip_set *set, void *value, const struct ip_set_ext *ext, INIT_LIST_HEAD(&e->list); list_set_init_extensions(set, ext, e); if (n) - list_set_replace(e, n); + list_set_replace(set, e, n); else if (next) list_add_tail_rcu(&e->list, &next->list); else if (prev) @@ -486,6 +490,7 @@ list_set_list(const struct ip_set *set, const struct list_set *map = set->data; struct nlattr *atd, *nested; u32 i = 0, first = cb->args[IPSET_CB_ARG0]; + char name[IPSET_MAXNAMELEN]; struct set_elem *e; int ret = 0; @@ -504,8 +509,8 @@ list_set_list(const struct ip_set *set, nested = ipset_nest_start(skb, IPSET_ATTR_DATA); if (!nested) goto nla_put_failure; - if (nla_put_string(skb, IPSET_ATTR_NAME, - ip_set_name_byindex(map->net, e->id))) + ip_set_name_byindex(map->net, e->id, name); + if (nla_put_string(skb, IPSET_ATTR_NAME, name)) goto nla_put_failure; if (ip_set_put_extensions(skb, set, e, true)) goto nla_put_failure;