Skip to contentSkip to conversation

Netfilter fixes for net

PATCH In review Pablo Neira Ayuso 2 participants 7 messages Aug 4, 2020

Subsystems (4)
lore
Share & tools
lore

Pablo Neira Ayuso

author maintainer
#0 lore raw

cover Netfilter fixes for net

Hi,

The following patchset contains Netfilter fixes for net:

1) Flush the cleanup xtables worker to make sure destructors have completed, from Florian Westphal.

2) iifgroup is matching erroneously, also from Florian.

3) Add selftest for meta interface matching, from Florian Westphal.

4) Move nf_ct_offload_timeout() to header, from Roi Dayan.

5) Call nf_ct_offload_timeout() from flow_offload_add() to make sure garbage collection does not evict offloaded flow, from Roi Dayan.

Please, pull these changes from:

git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git

Thank you!

----------------------------------------------------------------

The following changes since commit 85496a29224188051b6135eb38da8afd4c584765:

net: gemini: Fix missing clk_disable_unprepare() in error path of gemini_ethernet_port_probe() (2020-07-30 17:45:13 -0700)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git HEAD

for you to fetch changes up to 4203b19c27967d9eff6928f6a733f81892ffc592:

netfilter: flowtable: Set offload timeout when adding flow (2020-08-03 12:37:24 +0200)

----------------------------------------------------------------
Florian Westphal (3):
netfilter: nft_compat: make sure xtables destructors have run
netfilter: nft_meta: fix iifgroup matching
selftests: netfilter: add meta iif/oif match test

Roi Dayan (2):
netfilter: conntrack: Move nf_ct_offload_timeout to header file
netfilter: flowtable: Set offload timeout when adding flow

 include/net/netfilter/nf_conntrack.h          |  12 +++
 include/net/netfilter/nf_tables.h             |   2 +
 net/netfilter/nf_conntrack_core.c             |  12 ---
 net/netfilter/nf_flow_table_core.c            |   2 +
 net/netfilter/nf_tables_api.c                 |  10 ++-
 net/netfilter/nft_compat.c                    |  36 +++++++-
 net/netfilter/nft_meta.c                      |   2 +-
 tools/testing/selftests/netfilter/Makefile    |   2 +-
 tools/testing/selftests/netfilter/nft_meta.sh | 124 ++++++++++++++++++++++++++
 9 files changed, 182 insertions(+), 20 deletions(-)
 create mode 100755 tools/testing/selftests/netfilter/nft_meta.sh
author maintainer
#1 lore raw

PATCH 1/5 netfilter: nft_compat: make sure xtables destructors have run

From: Florian Westphal <[email protected]>

Pablo Neira found that after recent update of xt_IDLETIMER the iptables-nft tests sometimes show an error.

He tracked this down to the delayed cleanup used by nf_tables core:
del rule (transaction A)
add rule (transaction B)

Its possible that by time transaction B (both in same netns) runs, the xt target destructor has not been invoked yet.

For native nft expressions this is no problem because all expressions that have such side effects make sure these are handled from the commit phase, rather than async cleanup.

For nft_compat however this isn't true.

Instead of forcing synchronous behaviour for nft_compat, keep track of the number of outstanding destructor calls.

When we attempt to create a new expression, flush the cleanup worker to make sure destructors have completed.

With lots of help from Pablo Neira.

---
 include/net/netfilter/nf_tables.h |  2 ++
 net/netfilter/nf_tables_api.c     | 10 +++++++--
 net/netfilter/nft_compat.c        | 36 +++++++++++++++++++++++++++----
 3 files changed, 42 insertions(+), 6 deletions(-)
include/net/netfilter/nf_tables.h +2 -0
index 6f0f6fca9ac3..2571c09be8bb 100644
@@ -1498,4 +1498,6 @@ void nft_chain_filter_fini(void);
 
 void __init nft_chain_route_init(void);
 void nft_chain_route_fini(void);
+
+void nf_tables_trans_destroy_flush_work(void);
 #endif /* _NET_NF_TABLES_H */
net/netfilter/nf_tables_api.c +8 -2
index 88325b264737..79e4db3cadec 100644
@@ -7290,6 +7290,12 @@ static void nf_tables_trans_destroy_work(struct work_struct *w)
 	}
 }
 
+void nf_tables_trans_destroy_flush_work(void)
+{
+	flush_work(&trans_destroy_work);
+}
+EXPORT_SYMBOL_GPL(nf_tables_trans_destroy_flush_work);
+
 static int nf_tables_commit_chain_prepare(struct net *net, struct nft_chain *chain)
 {
 	struct nft_rule *rule;
@@ -7472,9 +7478,9 @@ static void nf_tables_commit_release(struct net *net)
 	spin_unlock(&nf_tables_destroy_list_lock);
 
 	nf_tables_module_autoload_cleanup(net);
-	mutex_unlock(&net->nft.commit_mutex);
-
 	schedule_work(&trans_destroy_work);
+
+	mutex_unlock(&net->nft.commit_mutex);
 }
 
 static int nf_tables_commit(struct net *net, struct sk_buff *skb)
net/netfilter/nft_compat.c +32 -4
index aa1a066cb74b..6428856ccbec 100644
@@ -27,6 +27,8 @@ struct nft_xt_match_priv {
 	void *info;
 };
 
+static refcount_t nft_compat_pending_destroy = REFCOUNT_INIT(1);
+
 static int nft_compat_chain_validate_dependency(const struct nft_ctx *ctx,
 						const char *tablename)
 {
@@ -236,6 +238,15 @@ nft_target_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 
 	nft_target_set_tgchk_param(&par, ctx, target, info, &e, proto, inv);
 
+	/* xtables matches or targets can have side effects, e.g.
+	 * creation/destruction of /proc files.
+	 * The xt ->destroy functions are run asynchronously from
+	 * work queue.  If we have pending invocations we thus
+	 * need to wait for those to finish.
+	 */
+	if (refcount_read(&nft_compat_pending_destroy) > 1)
+		nf_tables_trans_destroy_flush_work();
+
 	ret = xt_check_target(&par, size, proto, inv);
 	if (ret < 0)
 		return ret;
@@ -247,6 +258,13 @@ nft_target_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 	return 0;
 }
 
+static void __nft_mt_tg_destroy(struct module *me, const struct nft_expr *expr)
+{
+	refcount_dec(&nft_compat_pending_destroy);
+	module_put(me);
+	kfree(expr->ops);
+}
+
 static void
 nft_target_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr)
 {
@@ -262,8 +280,7 @@ nft_target_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr)
 	if (par.target->destroy != NULL)
 		par.target->destroy(&par);
 
-	module_put(me);
-	kfree(expr->ops);
+	__nft_mt_tg_destroy(me, expr);
 }
 
 static int nft_extension_dump_info(struct sk_buff *skb, int attr,
@@ -494,8 +511,7 @@ __nft_match_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr,
 	if (par.match->destroy != NULL)
 		par.match->destroy(&par);
 
-	module_put(me);
-	kfree(expr->ops);
+	__nft_mt_tg_destroy(me, expr);
 }
 
 static void
@@ -700,6 +716,14 @@ static const struct nfnetlink_subsystem nfnl_compat_subsys = {
 
 static struct nft_expr_type nft_match_type;
 
+static void nft_mt_tg_deactivate(const struct nft_ctx *ctx,
+				 const struct nft_expr *expr,
+				 enum nft_trans_phase phase)
+{
+	if (phase == NFT_TRANS_COMMIT)
+		refcount_inc(&nft_compat_pending_destroy);
+}
+
 static const struct nft_expr_ops *
 nft_match_select_ops(const struct nft_ctx *ctx,
 		     const struct nlattr * const tb[])
@@ -738,6 +762,7 @@ nft_match_select_ops(const struct nft_ctx *ctx,
 	ops->type = &nft_match_type;
 	ops->eval = nft_match_eval;
 	ops->init = nft_match_init;
+	ops->deactivate = nft_mt_tg_deactivate,
 	ops->destroy = nft_match_destroy;
 	ops->dump = nft_match_dump;
 	ops->validate = nft_match_validate;
@@ -828,6 +853,7 @@ nft_target_select_ops(const struct nft_ctx *ctx,
 	ops->size = NFT_EXPR_SIZE(XT_ALIGN(target->targetsize));
 	ops->init = nft_target_init;
 	ops->destroy = nft_target_destroy;
+	ops->deactivate = nft_mt_tg_deactivate,
 	ops->dump = nft_target_dump;
 	ops->validate = nft_target_validate;
 	ops->data = target;
@@ -891,6 +917,8 @@ static void __exit nft_compat_module_exit(void)
 	nfnetlink_subsys_unregister(&nfnl_compat_subsys);
 	nft_unregister_expr(&nft_target_type);
 	nft_unregister_expr(&nft_match_type);
+
+	WARN_ON_ONCE(refcount_read(&nft_compat_pending_destroy) != 1);
 }
 
 MODULE_ALIAS_NFNL_SUBSYS(NFNL_SUBSYS_NFT_COMPAT);
Signature · 1 line
2.20.1
author maintainer
#2 lore raw

PATCH 2/5 netfilter: nft_meta: fix iifgroup matching

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

iifgroup matching erroneously checks the output interface.

Fixes 8724e819cc9a ("netfilter: nft_meta: move all interface related keys to helper")
---
 net/netfilter/nft_meta.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/netfilter/nft_meta.c +1 -1
index 951b6e87ed5d..7bc6537f3ccb 100644
@@ -253,7 +253,7 @@ static bool nft_meta_get_eval_ifname(enum nft_meta_keys key, u32 *dest,
 			return false;
 		break;
 	case NFT_META_IIFGROUP:
-		if (!nft_meta_store_ifgroup(dest, nft_out(pkt)))
+		if (!nft_meta_store_ifgroup(dest, nft_in(pkt)))
 			return false;
 		break;
 	case NFT_META_OIFGROUP:
Signature · 1 line
2.20.1
author maintainer
#3 lore raw

PATCH 3/5 selftests: netfilter: add meta iif/oif match test

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

simple test case, but would have caught this:

FAIL: iifgroupcount, want "packets 2", got

table inet filter {
        counter iifgroupcount {
                packets 0 bytes 0
        }
}
---
 tools/testing/selftests/netfilter/Makefile    |   2 +-
 tools/testing/selftests/netfilter/nft_meta.sh | 124 ++++++++++++++++++
 2 files changed, 125 insertions(+), 1 deletion(-)
 create mode 100755 tools/testing/selftests/netfilter/nft_meta.sh
tools/testing/selftests/netfilter/Makefile +1 -1
index a179f0dca8ce..a374e10ef506 100644
@@ -4,7 +4,7 @@
 TEST_PROGS := nft_trans_stress.sh nft_nat.sh bridge_brouter.sh \
 	conntrack_icmp_related.sh nft_flowtable.sh ipvs.sh \
 	nft_concat_range.sh nft_conntrack_helper.sh \
-	nft_queue.sh
+	nft_queue.sh nft_meta.sh
 
 LDLIBS = -lmnl
 TEST_GEN_FILES =  nf-queue
tools/testing/selftests/netfilter/nft_meta.sh new file +124 -0

124 lines not on this page · raw message

Signature · 1 line
2.20.1
author maintainer
#4 lore raw

PATCH 4/5 netfilter: conntrack: Move nf_ct_offload_timeout to header file

In reply to Pablo Neira Ayuso

From: Roi Dayan <[email protected]>

To be used by callers from other modules.

[ Rename DAY to NF_CT_DAY to avoid possible symbol name pollution
issue --Pablo ]

---
 include/net/netfilter/nf_conntrack.h | 12 ++++++++++++
 net/netfilter/nf_conntrack_core.c    | 12 ------------
 2 files changed, 12 insertions(+), 12 deletions(-)
include/net/netfilter/nf_conntrack.h +12 -0
index 90690e37a56f..c7bfddfc65b0 100644
@@ -279,6 +279,18 @@ static inline bool nf_ct_should_gc(const struct nf_conn *ct)
 	       !nf_ct_is_dying(ct);
 }
 
+#define	NF_CT_DAY	(86400 * HZ)
+
+/* Set an arbitrary timeout large enough not to ever expire, this save
+ * us a check for the IPS_OFFLOAD_BIT from the packet path via
+ * nf_ct_is_expired().
+ */
+static inline void nf_ct_offload_timeout(struct nf_conn *ct)
+{
+	if (nf_ct_expires(ct) < NF_CT_DAY / 2)
+		ct->timeout = nfct_time_stamp + NF_CT_DAY;
+}
+
 struct kernel_param;
 
 int nf_conntrack_set_hashsize(const char *val, const struct kernel_param *kp);
net/netfilter/nf_conntrack_core.c +0 -12
index f33d72c5b06e..c3cea50d1bcb 100644
@@ -1344,18 +1344,6 @@ static bool gc_worker_can_early_drop(const struct nf_conn *ct)
 	return false;
 }
 
-#define	DAY	(86400 * HZ)
-
-/* Set an arbitrary timeout large enough not to ever expire, this save
- * us a check for the IPS_OFFLOAD_BIT from the packet path via
- * nf_ct_is_expired().
- */
-static void nf_ct_offload_timeout(struct nf_conn *ct)
-{
-	if (nf_ct_expires(ct) < DAY / 2)
-		ct->timeout = nfct_time_stamp + DAY;
-}
-
 static void gc_worker(struct work_struct *work)
 {
 	unsigned int min_interval = max(HZ / GC_MAX_BUCKETS_DIV, 1u);
Signature · 1 line
2.20.1
author maintainer
#5 lore raw

PATCH 5/5 netfilter: flowtable: Set offload timeout when adding flow

In reply to Pablo Neira Ayuso

From: Roi Dayan <[email protected]>

On heavily loaded systems the GC can take time to go over all existing conns and reset their timeout. At that time other calls like from nf_conntrack_in() can call of nf_ct_is_expired() and see the conn as expired. To fix this when we set the offload bit we should also reset the timeout instead of counting on GC to finish first iteration over all conns before the initial timeout.

Fixes 90964016e5d3 ("netfilter: nf_conntrack: add IPS_OFFLOAD status bit")
---
 net/netfilter/nf_flow_table_core.c | 2 ++
 1 file changed, 2 insertions(+)
net/netfilter/nf_flow_table_core.c +2 -0
index b1eb5272b379..4f7a567c536e 100644
@@ -243,6 +243,8 @@ int flow_offload_add(struct nf_flowtable *flow_table, struct flow_offload *flow)
 		return err;
 	}
 
+	nf_ct_offload_timeout(flow->ct);
+
 	if (nf_flowtable_hw_offload(flow_table)) {
 		__set_bit(NF_FLOW_HW, &flow->flags);
 		nf_flow_offload_add(flow_table, flow);
Signature · 1 line
2.20.1
maintainer
#6 lore raw
In reply to Pablo Neira Ayuso

From: Pablo Neira Ayuso <[email protected]>
Date: Tue, 4 Aug 2020 22:02:03 +0200

Quoted 11 lines

The following patchset contains Netfilter fixes for net:

1) Flush the cleanup xtables worker to make sure destructors have completed, from Florian Westphal.

2) iifgroup is matching erroneously, also from Florian.

3) Add selftest for meta interface matching, from Florian Westphal.

4) Move nf_ct_offload_timeout() to header, from Roi Dayan.

Quoted 7 lines

5) Call nf_ct_offload_timeout() from flow_offload_add() to make sure garbage collection does not evict offloaded flow, from Roi Dayan.

Please, pull these changes from:

git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git

Pulled, thanks Pablo.