Skip to contentSkip to conversation

Netfilter/IPVS updates for net-next

PATCH New Pablo Neira Ayuso 2 participants 20 messages Jul 20, 2022 – Jul 21, 2022

Subsystems (6)
lore
Share & tools
lore

Pablo Neira Ayuso

author maintainer
#0 lore raw

cover Netfilter/IPVS updates for net-next

Hi,

The following patchset contains Netfilter/IPVS updates for net-next:

1) Simplify nf_ct_get_tuple(), from Jackie Liu.

2) Add format to request_module() call, from Bill Wendling.

3) Add /proc/net/stats/nf_flowtable to monitor in-flight pending hardware offload objects to be processed, from Vlad Buslov.

4) Missing rcu annotation and accessors in the netfilter tree, from Florian Westphal.

5) Merge h323 conntrack helper nat hooks into single object, also from Florian.

6) A batch of update to fix sparse warnings treewide, from Florian Westphal.

7) Move nft_cmp_fast_mask() where it used, from Florian.

8) Missing const in nf_nat_initialized(), from James Yonan.

9) Use bitmap API for Maglev IPVS scheduler, from Christophe Jaillet.

10) Use refcount_inc instead of _inc_not_zero in flowtable, from Florian Westphal.

11) Remove pr_debug in xt_TPROXY, from Nathan Cancellor.

Please, pull these changes from:

git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git

Thanks.

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

The following changes since commit edb2c3476db9898a63fb5d0011ecaa43ebf46c9b:

fddi/skfp: fix repeated words in comments (2022-07-11 14:12:54 +0100)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git HEAD

for you to fetch changes up to aa8c7cdbae58b695ed79a0129b6b8c887b25969f:

netfilter: xt_TPROXY: remove pr_debug invocations (2022-07-21 00:56:00 +0200)

----------------------------------------------------------------
Bill Wendling (1):
netfilter: conntrack: use correct format characters

Christophe JAILLET (1):
netfilter: ipvs: Use the bitmap API to allocate bitmaps

Florian Westphal (10):
netfilter: nf_conntrack: add missing __rcu annotations
netfilter: nf_conntrack: use rcu accessors where needed
netfilter: h323: merge nat hook pointers into one
netfilter: nfnetlink: add missing __be16 cast
netfilter: x_tables: use correct integer types
netfilter: nf_tables: use the correct get/put helpers
netfilter: nf_tables: add and use BE register load-store helpers
netfilter: nf_tables: use correct integer types
netfilter: nf_tables: move nft_cmp_fast_mask to where its used
netfilter: flowtable: prefer refcount_inc

Jackie Liu (1):
netfilter: conntrack: use fallthrough to cleanup

James Yonan (1):
netfilter: nf_nat: in nf_nat_initialized(), use const struct nf_conn *

Justin Stitt (1):
netfilter: xt_TPROXY: remove pr_debug invocations

Vlad Buslov (2):
net/sched: act_ct: set 'net' pointer when creating new nf_flow_table
netfilter: nf_flow_table: count pending offload workqueue tasks

Zhang Jiaming (1):
netfilter: nft_set_bitmap: Fix spelling mistake

 include/linux/netfilter/nf_conntrack_h323.h  | 109 +++++------
 include/linux/netfilter/nf_conntrack_sip.h   |   2 +-
 include/net/net_namespace.h                  |   6 +
 include/net/netfilter/nf_conntrack_timeout.h |   2 +-
 include/net/netfilter/nf_flow_table.h        |  21 +++
 include/net/netfilter/nf_nat.h               |   2 +-
 include/net/netfilter/nf_tables.h            |  15 ++
 include/net/netfilter/nf_tables_core.h       |  10 --
 include/net/netns/flow_table.h               |  14 ++
 net/bridge/netfilter/nft_meta_bridge.c       |   2 +-
 net/ipv4/netfilter/nf_nat_h323.c             |  42 ++---
 net/netfilter/Kconfig                        |   9 +
 net/netfilter/Makefile                       |   1 +
 net/netfilter/ipvs/ip_vs_mh.c                |   5 +-
 net/netfilter/nf_conntrack_broadcast.c       |   6 +-
 net/netfilter/nf_conntrack_core.c            |   8 +-
 net/netfilter/nf_conntrack_h323_main.c       | 260 ++++++++++-----------------
 net/netfilter/nf_conntrack_helper.c          |   4 +-
 net/netfilter/nf_conntrack_netlink.c         |   9 +-
 net/netfilter/nf_conntrack_pptp.c            |   2 +-
 net/netfilter/nf_conntrack_sip.c             |   9 +-
 net/netfilter/nf_conntrack_timeout.c         |  18 +-
 net/netfilter/nf_flow_table_core.c           |  73 +++++++-
 net/netfilter/nf_flow_table_offload.c        |  17 +-
 net/netfilter/nf_flow_table_procfs.c         |  80 +++++++++
 net/netfilter/nfnetlink.c                    |   2 +-
 net/netfilter/nfnetlink_cthelper.c           |  10 +-
 net/netfilter/nft_byteorder.c                |   3 +-
 net/netfilter/nft_cmp.c                      |  18 +-
 net/netfilter/nft_ct.c                       |   4 +-
 net/netfilter/nft_exthdr.c                   |  10 +-
 net/netfilter/nft_osf.c                      |   2 +-
 net/netfilter/nft_set_bitmap.c               |   4 +-
 net/netfilter/nft_socket.c                   |   8 +-
 net/netfilter/nft_tproxy.c                   |   6 +-
 net/netfilter/nft_tunnel.c                   |   3 +-
 net/netfilter/nft_xfrm.c                     |   8 +-
 net/netfilter/xt_CT.c                        |  23 ++-
 net/netfilter/xt_DSCP.c                      |   8 +-
 net/netfilter/xt_TCPMSS.c                    |   4 +-
 net/netfilter/xt_TPROXY.c                    |  25 +--
 net/netfilter/xt_connlimit.c                 |   6 +-
 net/sched/act_ct.c                           |   5 +-
 43 files changed, 518 insertions(+), 357 deletions(-)
 create mode 100644 include/net/netns/flow_table.h
 create mode 100644 net/netfilter/nf_flow_table_procfs.c
author maintainer
#1 lore raw

PATCH 1/18 netfilter: conntrack: use fallthrough to cleanup

From: Jackie Liu <[email protected]>

These cases all use the same function. we can simplify the code through fallthrough.

$ size net/netfilter/nf_conntrack_core.o
        text	   data	    bss	    dec	    hex	filename
before  81601	  81430	    768	 163799	  27fd7	net/netfilter/nf_conntrack_core.o
after   80361	  81430	    768	 162559	  27aff	net/netfilter/nf_conntrack_core.o

Arch: aarch64
Gcc : gcc version 9.4.0 (Ubuntu 9.4.0-1ubuntu1~20.04.1)

---
 net/netfilter/nf_conntrack_core.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)
net/netfilter/nf_conntrack_core.c +3 -5
index 082a2fd8d85b..363c3e8f9d1b 100644
@@ -329,20 +329,18 @@ nf_ct_get_tuple(const struct sk_buff *skb,
 		return gre_pkt_to_tuple(skb, dataoff, net, tuple);
 #endif
 	case IPPROTO_TCP:
-	case IPPROTO_UDP: /* fallthrough */
-		return nf_ct_get_tuple_ports(skb, dataoff, tuple);
+	case IPPROTO_UDP:
 #ifdef CONFIG_NF_CT_PROTO_UDPLITE
 	case IPPROTO_UDPLITE:
-		return nf_ct_get_tuple_ports(skb, dataoff, tuple);
 #endif
 #ifdef CONFIG_NF_CT_PROTO_SCTP
 	case IPPROTO_SCTP:
-		return nf_ct_get_tuple_ports(skb, dataoff, tuple);
 #endif
 #ifdef CONFIG_NF_CT_PROTO_DCCP
 	case IPPROTO_DCCP:
-		return nf_ct_get_tuple_ports(skb, dataoff, tuple);
 #endif
+		/* fallthrough */
+		return nf_ct_get_tuple_ports(skb, dataoff, tuple);
 	default:
 		break;
 	}
Signature · 1 line
2.30.2
#2 lore raw

Re: [PATCH nf-next 01/18] netfilter: conntrack: use fallthrough to cleanup

This series was applied to netdev/net-next.git (master) by Pablo Neira Ayuso <[email protected]>:

Hello:

This series was applied to netdev/net-next.git (master) by Pablo Neira Ayuso <[email protected]>:

Quoted 8 lines · On Thu, 21 Jul 2022 01:07:37 +0200 you wrote:

From: Jackie Liu <[email protected]>

These cases all use the same function. we can simplify the code through fallthrough.

$ size net/netfilter/nf_conntrack_core.o

[...]

Here is the summary with links:
- [nf-next,01/18] netfilter: conntrack: use fallthrough to cleanup

    https://git.kernel.org/netdev/net-next/c/6be791561212
  - [nf-next,02/18] netfilter: conntrack: use correct format characters
    https://git.kernel.org/netdev/net-next/c/b8acd43148c0
  - [nf-next,03/18] net/sched: act_ct: set 'net' pointer when creating new nf_flow_table
    https://git.kernel.org/netdev/net-next/c/fc54d9065f90
  - [nf-next,04/18] netfilter: nf_flow_table: count pending offload workqueue tasks
    https://git.kernel.org/netdev/net-next/c/b038177636f8
  - [nf-next,05/18] netfilter: nf_conntrack: add missing __rcu annotations
    https://git.kernel.org/netdev/net-next/c/6976890e8998
  - [nf-next,06/18] netfilter: nf_conntrack: use rcu accessors where needed
    https://git.kernel.org/netdev/net-next/c/e14575fa7529
  - [nf-next,07/18] netfilter: h323: merge nat hook pointers into one
    https://git.kernel.org/netdev/net-next/c/d3f2d0a292c2
  - [nf-next,08/18] netfilter: nft_set_bitmap: Fix spelling mistake
    https://git.kernel.org/netdev/net-next/c/f72547473fcd
  - [nf-next,09/18] netfilter: nfnetlink: add missing __be16 cast
    https://git.kernel.org/netdev/net-next/c/ec6f2ff0a398
  - [nf-next,10/18] netfilter: x_tables: use correct integer types
    https://git.kernel.org/netdev/net-next/c/168141f7e0b4
  - [nf-next,11/18] netfilter: nf_tables: use the correct get/put helpers
    https://git.kernel.org/netdev/net-next/c/d86473bf2ff3
  - [nf-next,12/18] netfilter: nf_tables: add and use BE register load-store helpers
    https://git.kernel.org/netdev/net-next/c/7278b3c1e4eb
  - [nf-next,13/18] netfilter: nf_tables: use correct integer types
    https://git.kernel.org/netdev/net-next/c/ffb3d9a30cc6
  - [nf-next,14/18] netfilter: nf_tables: move nft_cmp_fast_mask to where its used
    https://git.kernel.org/netdev/net-next/c/6b77205374fd
  - [nf-next,15/18] netfilter: nf_nat: in nf_nat_initialized(), use const struct nf_conn *
    https://git.kernel.org/netdev/net-next/c/9d2f00fb0a0c
  - [nf-next,16/18] netfilter: ipvs: Use the bitmap API to allocate bitmaps
    https://git.kernel.org/netdev/net-next/c/5787db7c9053
  - [nf-next,17/18] netfilter: flowtable: prefer refcount_inc
    https://git.kernel.org/netdev/net-next/c/f02e7dc4cff8

- [nf-next,18/18] netfilter: xt_TPROXY: remove pr_debug invocations (no matching commit)

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
author maintainer
#3 lore raw

PATCH 2/18 netfilter: conntrack: use correct format characters

In reply to Pablo Neira Ayuso

From: Bill Wendling <[email protected]>

When compiling with -Wformat, clang emits the following warnings:

net/netfilter/nf_conntrack_helper.c:168:18: error: format string is not a string literal (potentially insecure) [-Werror,-Wformat-security]

                request_module(mod_name);
                               ^~~~~~~~

Use a string literal for the format string.

---
 net/netfilter/nf_conntrack_helper.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/netfilter/nf_conntrack_helper.c +1 -1
index c12a87ebc3ee..1e0424d37abc 100644
@@ -165,7 +165,7 @@ nf_nat_helper_try_module_get(const char *name, u16 l3num, u8 protonum)
 	if (!nat) {
 		snprintf(mod_name, sizeof(mod_name), "%s", h->nat_mod_name);
 		rcu_read_unlock();
-		request_module(mod_name);
+		request_module("%s", mod_name);
 
 		rcu_read_lock();
 		nat = nf_conntrack_nat_helper_find(mod_name);
Signature · 1 line
2.30.2
author maintainer
#4 lore raw

PATCH 3/18 net/sched: act_ct: set 'net' pointer when creating new nf_flow_table

In reply to Pablo Neira Ayuso

From: Vlad Buslov <[email protected]>

Following patches in series use the pointer to access flow table offload debug variables.

---
 net/sched/act_ct.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/sched/act_ct.c +3 -2
index e013253b10d1..d55afb8d14be 100644
@@ -277,7 +277,7 @@ static struct nf_flowtable_type flowtable_ct = {
 	.owner		= THIS_MODULE,
 };
 
-static int tcf_ct_flow_table_get(struct tcf_ct_params *params)
+static int tcf_ct_flow_table_get(struct net *net, struct tcf_ct_params *params)
 {
 	struct tcf_ct_flow_table *ct_ft;
 	int err = -ENOMEM;
@@ -303,6 +303,7 @@ static int tcf_ct_flow_table_get(struct tcf_ct_params *params)
 	err = nf_flow_table_init(&ct_ft->nf_ft);
 	if (err)
 		goto err_init;
+	write_pnet(&ct_ft->nf_ft.net, net);
 
 	__module_get(THIS_MODULE);
 out_unlock:
@@ -1391,7 +1392,7 @@ static int tcf_ct_init(struct net *net, struct nlattr *nla,
 	if (err)
 		goto cleanup;
 
-	err = tcf_ct_flow_table_get(params);
+	err = tcf_ct_flow_table_get(net, params);
 	if (err)
 		goto cleanup;
 
Signature · 1 line
2.30.2
author maintainer
#5 lore raw

PATCH 4/18 netfilter: nf_flow_table: count pending offload workqueue tasks

In reply to Pablo Neira Ayuso

From: Vlad Buslov <[email protected]>

To improve hardware offload debuggability count pending 'add', 'del' and 'stats' flow_table offload workqueue tasks. Counters are incremented before scheduling new task and decremented when workqueue handler finishes executing. These counters allow user to diagnose congestion on hardware offload workqueues that can happen when either CPU is starved and workqueue jobs are executed at lower rate than new ones are added or when hardware/driver can't keep up with the rate.

Implement the described counters as percpu counters inside new struct netns_ft which is stored inside struct net. Expose them via new procfs file '/proc/net/stats/nf_flowtable' that is similar to existing 'nf_conntrack' file.

---
 include/net/net_namespace.h           |  6 ++
 include/net/netfilter/nf_flow_table.h | 21 +++++++
 include/net/netns/flow_table.h        | 14 +++++
 net/netfilter/Kconfig                 |  9 +++
 net/netfilter/Makefile                |  1 +
 net/netfilter/nf_flow_table_core.c    | 62 ++++++++++++++++++++-
 net/netfilter/nf_flow_table_offload.c | 17 +++++-
 net/netfilter/nf_flow_table_procfs.c  | 80 +++++++++++++++++++++++++++
 8 files changed, 206 insertions(+), 4 deletions(-)
 create mode 100644 include/net/netns/flow_table.h
 create mode 100644 net/netfilter/nf_flow_table_procfs.c
include/net/net_namespace.h +6 -0
index 20a2992901c2..8c3587d5c308 100644
@@ -26,6 +26,9 @@
 #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
 #include <net/netns/conntrack.h>
 #endif
+#if IS_ENABLED(CONFIG_NF_FLOW_TABLE)
+#include <net/netns/flow_table.h>
+#endif
 #include <net/netns/nftables.h>
 #include <net/netns/xfrm.h>
 #include <net/netns/mpls.h>
@@ -142,6 +145,9 @@ struct net {
 #if defined(CONFIG_NF_TABLES) || defined(CONFIG_NF_TABLES_MODULE)
 	struct netns_nftables	nft;
 #endif
+#if IS_ENABLED(CONFIG_NF_FLOW_TABLE)
+	struct netns_ft ft;
+#endif
 #endif
 #ifdef CONFIG_WEXT_CORE
 	struct sk_buff_head	wext_nlevents;
include/net/netfilter/nf_flow_table.h +21 -0
index 64daafd1fc41..d5326c44b453 100644
@@ -335,4 +335,25 @@ static inline __be16 nf_flow_pppoe_proto(const struct sk_buff *skb)
 	return 0;
 }
 
+#define NF_FLOW_TABLE_STAT_INC(net, count) __this_cpu_inc((net)->ft.stat->count)
+#define NF_FLOW_TABLE_STAT_DEC(net, count) __this_cpu_dec((net)->ft.stat->count)
+#define NF_FLOW_TABLE_STAT_INC_ATOMIC(net, count)	\
+	this_cpu_inc((net)->ft.stat->count)
+#define NF_FLOW_TABLE_STAT_DEC_ATOMIC(net, count)	\
+	this_cpu_dec((net)->ft.stat->count)
+
+#ifdef CONFIG_NF_FLOW_TABLE_PROCFS
+int nf_flow_table_init_proc(struct net *net);
+void nf_flow_table_fini_proc(struct net *net);
+#else
+static inline int nf_flow_table_init_proc(struct net *net)
+{
+	return 0;
+}
+
+static inline void nf_flow_table_fini_proc(struct net *net)
+{
+}
+#endif /* CONFIG_NF_FLOW_TABLE_PROCFS */
+
 #endif /* _NF_FLOW_TABLE_H */
include/net/netns/flow_table.h new file +14 -0
new file mode 100644
index 000000000000..1c5fc657e267
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __NETNS_FLOW_TABLE_H
+#define __NETNS_FLOW_TABLE_H
+
+struct nf_flow_table_stat {
+	unsigned int count_wq_add;
+	unsigned int count_wq_del;
+	unsigned int count_wq_stats;
+};
+
+struct netns_ft {
+	struct nf_flow_table_stat __percpu *stat;
+};
+#endif
net/netfilter/Kconfig +9 -0
index ddc54b6d18ee..df6abbfe0079 100644
@@ -734,6 +734,15 @@ config NF_FLOW_TABLE
 
 	  To compile it as a module, choose M here.
 
+config NF_FLOW_TABLE_PROCFS
+	bool "Supply flow table statistics in procfs"
+	default y
+	depends on PROC_FS
+	depends on SYSCTL
+	help
+	  This option enables for the flow table offload statistics
+	  to be shown in procfs under net/netfilter/nf_flowtable.
+
 config NETFILTER_XTABLES
 	tristate "Netfilter Xtables support (required for ip_tables)"
 	default m if NETFILTER_ADVANCED=n
net/netfilter/Makefile +1 -0
index 238b6a620e88..06df49ea6329 100644
@@ -128,6 +128,7 @@ obj-$(CONFIG_NFT_FWD_NETDEV)	+= nft_fwd_netdev.o
 obj-$(CONFIG_NF_FLOW_TABLE)	+= nf_flow_table.o
 nf_flow_table-objs		:= nf_flow_table_core.o nf_flow_table_ip.o \
 				   nf_flow_table_offload.o
+nf_flow_table-$(CONFIG_NF_FLOW_TABLE_PROCFS) += nf_flow_table_procfs.o
 
 obj-$(CONFIG_NF_FLOW_TABLE_INET) += nf_flow_table_inet.o
 
net/netfilter/nf_flow_table_core.c +61 -1

75 lines not on this page · raw message

net/netfilter/nf_flow_table_offload.c +14 -3
index 11b6e1942092..103b6cbf257f 100644
@@ -967,17 +967,22 @@ static void flow_offload_work_stats(struct flow_offload_work *offload)
 static void flow_offload_work_handler(struct work_struct *work)
 {
 	struct flow_offload_work *offload;
+	struct net *net;
 
 	offload = container_of(work, struct flow_offload_work, work);
+	net = read_pnet(&offload->flowtable->net);
 	switch (offload->cmd) {
 		case FLOW_CLS_REPLACE:
 			flow_offload_work_add(offload);
+			NF_FLOW_TABLE_STAT_DEC_ATOMIC(net, count_wq_add);
 			break;
 		case FLOW_CLS_DESTROY:
 			flow_offload_work_del(offload);
+			NF_FLOW_TABLE_STAT_DEC_ATOMIC(net, count_wq_del);
 			break;
 		case FLOW_CLS_STATS:
 			flow_offload_work_stats(offload);
+			NF_FLOW_TABLE_STAT_DEC_ATOMIC(net, count_wq_stats);
 			break;
 		default:
 			WARN_ON_ONCE(1);
@@ -989,12 +994,18 @@ static void flow_offload_work_handler(struct work_struct *work)
 
 static void flow_offload_queue_work(struct flow_offload_work *offload)
 {
-	if (offload->cmd == FLOW_CLS_REPLACE)
+	struct net *net = read_pnet(&offload->flowtable->net);
+
+	if (offload->cmd == FLOW_CLS_REPLACE) {
+		NF_FLOW_TABLE_STAT_INC(net, count_wq_add);
 		queue_work(nf_flow_offload_add_wq, &offload->work);
-	else if (offload->cmd == FLOW_CLS_DESTROY)
+	} else if (offload->cmd == FLOW_CLS_DESTROY) {
+		NF_FLOW_TABLE_STAT_INC(net, count_wq_del);
 		queue_work(nf_flow_offload_del_wq, &offload->work);
-	else
+	} else {
+		NF_FLOW_TABLE_STAT_INC(net, count_wq_stats);
 		queue_work(nf_flow_offload_stats_wq, &offload->work);
+	}
 }
 
 static struct flow_offload_work *
net/netfilter/nf_flow_table_procfs.c new file +80 -0

80 lines not on this page · raw message

Signature · 1 line
2.30.2
author maintainer
#6 lore raw

PATCH 5/18 netfilter: nf_conntrack: add missing __rcu annotations

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Access to the hook pointers use correct helpers but the pointers lack the needed __rcu annotation.

---
 include/linux/netfilter/nf_conntrack_sip.h   | 2 +-
 include/net/netfilter/nf_conntrack_timeout.h | 2 +-
 net/netfilter/nf_conntrack_pptp.c            | 2 +-
 net/netfilter/nf_conntrack_sip.c             | 2 +-
 net/netfilter/nf_conntrack_timeout.c         | 2 +-
 5 files changed, 5 insertions(+), 5 deletions(-)
include/linux/netfilter/nf_conntrack_sip.h +1 -1
index c620521c42bc..dbc614dfe0d5 100644
@@ -164,7 +164,7 @@ struct nf_nat_sip_hooks {
 				  unsigned int medialen,
 				  union nf_inet_addr *rtp_addr);
 };
-extern const struct nf_nat_sip_hooks *nf_nat_sip_hooks;
+extern const struct nf_nat_sip_hooks __rcu *nf_nat_sip_hooks;
 
 int ct_sip_parse_request(const struct nf_conn *ct, const char *dptr,
 			 unsigned int datalen, unsigned int *matchoff,
include/net/netfilter/nf_conntrack_timeout.h +1 -1
index fea258983d23..9fdaba911de6 100644
@@ -105,7 +105,7 @@ struct nf_ct_timeout_hooks {
 	void (*timeout_put)(struct nf_ct_timeout *timeout);
 };
 
-extern const struct nf_ct_timeout_hooks *nf_ct_timeout_hook;
+extern const struct nf_ct_timeout_hooks __rcu *nf_ct_timeout_hook;
 #endif
 
 #endif /* _NF_CONNTRACK_TIMEOUT_H */
net/netfilter/nf_conntrack_pptp.c +1 -1
index f3fa367b455f..4c679638df06 100644
@@ -45,7 +45,7 @@ MODULE_ALIAS_NFCT_HELPER("pptp");
 
 static DEFINE_SPINLOCK(nf_pptp_lock);
 
-const struct nf_nat_pptp_hook *nf_nat_pptp_hook;
+const struct nf_nat_pptp_hook __rcu *nf_nat_pptp_hook;
 EXPORT_SYMBOL_GPL(nf_nat_pptp_hook);
 
 #if defined(DEBUG) || defined(CONFIG_DYNAMIC_DEBUG)
net/netfilter/nf_conntrack_sip.c +1 -1
index b83dc9bf0a5d..a88b43624b27 100644
@@ -60,7 +60,7 @@ module_param(sip_external_media, int, 0600);
 MODULE_PARM_DESC(sip_external_media, "Expect Media streams between external "
 				     "endpoints (default 0)");
 
-const struct nf_nat_sip_hooks *nf_nat_sip_hooks;
+const struct nf_nat_sip_hooks __rcu *nf_nat_sip_hooks;
 EXPORT_SYMBOL_GPL(nf_nat_sip_hooks);
 
 static int string_len(const struct nf_conn *ct, const char *dptr,
net/netfilter/nf_conntrack_timeout.c +1 -1
index 0f828d05ea60..821365ed5b2c 100644
@@ -22,7 +22,7 @@
 #include <net/netfilter/nf_conntrack_l4proto.h>
 #include <net/netfilter/nf_conntrack_timeout.h>
 
-const struct nf_ct_timeout_hooks *nf_ct_timeout_hook __read_mostly;
+const struct nf_ct_timeout_hooks __rcu *nf_ct_timeout_hook __read_mostly;
 EXPORT_SYMBOL_GPL(nf_ct_timeout_hook);
 
 static int untimeout(struct nf_conn *ct, void *timeout)
Signature · 1 line
2.30.2
author maintainer
#7 lore raw

PATCH 6/18 netfilter: nf_conntrack: use rcu accessors where needed

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Sparse complains about direct access to the 'helper' and timeout members. Both have __rcu annotation, so use the accessors.

xt_CT is fine, accesses occur before the structure is visible to other cpus. Switch to rcu accessors there as well to reduce noise.

---
 net/netfilter/nf_conntrack_broadcast.c |  6 +++++-
 net/netfilter/nf_conntrack_helper.c    |  2 +-
 net/netfilter/nf_conntrack_netlink.c   |  9 +++++++--
 net/netfilter/nf_conntrack_sip.c       |  7 ++++++-
 net/netfilter/nf_conntrack_timeout.c   | 16 +++++++++++++---
 net/netfilter/nfnetlink_cthelper.c     | 10 +++++++---
 net/netfilter/xt_CT.c                  | 23 ++++++++++++++++++-----
 7 files changed, 57 insertions(+), 16 deletions(-)
net/netfilter/nf_conntrack_broadcast.c +5 -1
index 1ba6becc3079..9fb9b8031298 100644
@@ -20,6 +20,7 @@ int nf_conntrack_broadcast_help(struct sk_buff *skb,
 				enum ip_conntrack_info ctinfo,
 				unsigned int timeout)
 {
+	const struct nf_conntrack_helper *helper;
 	struct nf_conntrack_expect *exp;
 	struct iphdr *iph = ip_hdr(skb);
 	struct rtable *rt = skb_rtable(skb);
@@ -58,7 +59,10 @@ int nf_conntrack_broadcast_help(struct sk_buff *skb,
 		goto out;
 
 	exp->tuple                = ct->tuplehash[IP_CT_DIR_REPLY].tuple;
-	exp->tuple.src.u.udp.port = help->helper->tuple.src.u.udp.port;
+
+	helper = rcu_dereference(help->helper);
+	if (helper)
+		exp->tuple.src.u.udp.port = helper->tuple.src.u.udp.port;
 
 	exp->mask.src.u3.ip       = mask;
 	exp->mask.src.u.udp.port  = htons(0xFFFF);
net/netfilter/nf_conntrack_helper.c +1 -1
index 1e0424d37abc..e96b32221444 100644
@@ -249,7 +249,7 @@ int __nf_ct_try_assign_helper(struct nf_conn *ct, struct nf_conn *tmpl,
 	if (tmpl != NULL) {
 		help = nfct_help(tmpl);
 		if (help != NULL) {
-			helper = help->helper;
+			helper = rcu_dereference(help->helper);
 			set_bit(IPS_HELPER_BIT, &ct->status);
 		}
 	}
net/netfilter/nf_conntrack_netlink.c +7 -2
index 722af5e309ba..25c2c0de78f0 100644
@@ -2004,7 +2004,7 @@ static int ctnetlink_change_helper(struct nf_conn *ct,
 	}
 
 	if (help) {
-		if (help->helper == helper) {
+		if (rcu_access_pointer(help->helper) == helper) {
 			/* update private helper data if allowed. */
 			if (helper->from_nlattr)
 				helper->from_nlattr(helpinfo, ct);
@@ -3412,12 +3412,17 @@ static int ctnetlink_get_expect(struct sk_buff *skb,
 
 static bool expect_iter_name(struct nf_conntrack_expect *exp, void *data)
 {
+	struct nf_conntrack_helper *helper;
 	const struct nf_conn_help *m_help;
 	const char *name = data;
 
 	m_help = nfct_help(exp->master);
 
-	return strcmp(m_help->helper->name, name) == 0;
+	helper = rcu_dereference(m_help->helper);
+	if (!helper)
+		return false;
+
+	return strcmp(helper->name, name) == 0;
 }
 
 static bool expect_iter_all(struct nf_conntrack_expect *exp, void *data)
net/netfilter/nf_conntrack_sip.c +6 -1
index a88b43624b27..daf06f71d31c 100644
@@ -1229,6 +1229,7 @@ static int process_register_request(struct sk_buff *skb, unsigned int protoff,
 	struct nf_conntrack_expect *exp;
 	union nf_inet_addr *saddr, daddr;
 	const struct nf_nat_sip_hooks *hooks;
+	struct nf_conntrack_helper *helper;
 	__be16 port;
 	u8 proto;
 	unsigned int expires = 0;
@@ -1289,10 +1290,14 @@ static int process_register_request(struct sk_buff *skb, unsigned int protoff,
 	if (sip_direct_signalling)
 		saddr = &ct->tuplehash[!dir].tuple.src.u3;
 
+	helper = rcu_dereference(nfct_help(ct)->helper);
+	if (!helper)
+		return NF_DROP;
+
 	nf_ct_expect_init(exp, SIP_EXPECT_SIGNALLING, nf_ct_l3num(ct),
 			  saddr, &daddr, proto, NULL, &port);
 	exp->timeout.expires = sip_timeout * HZ;
-	exp->helper = nfct_help(ct)->helper;
+	exp->helper = helper;
 	exp->flags = NF_CT_EXPECT_PERMANENT | NF_CT_EXPECT_INACTIVE;
 
 	hooks = rcu_dereference(nf_nat_sip_hooks);
net/netfilter/nf_conntrack_timeout.c +13 -3
index 821365ed5b2c..0cc584d3dbb1 100644
@@ -29,8 +29,14 @@ static int untimeout(struct nf_conn *ct, void *timeout)
 {
 	struct nf_conn_timeout *timeout_ext = nf_ct_timeout_find(ct);
 
-	if (timeout_ext && (!timeout || timeout_ext->timeout == timeout))
-		RCU_INIT_POINTER(timeout_ext->timeout, NULL);
+	if (timeout_ext) {
+		const struct nf_ct_timeout *t;
+
+		t = rcu_access_pointer(timeout_ext->timeout);
+
+		if (!timeout || t == timeout)
+			RCU_INIT_POINTER(timeout_ext->timeout, NULL);
+	}
 
 	/* We are not intended to delete this conntrack. */
 	return 0;
@@ -127,7 +133,11 @@ void nf_ct_destroy_timeout(struct nf_conn *ct)
 	if (h) {
 		timeout_ext = nf_ct_timeout_find(ct);
 		if (timeout_ext) {
-			h->timeout_put(timeout_ext->timeout);
+			struct nf_ct_timeout *t;
+
+			t = rcu_dereference(timeout_ext->timeout);
+			if (t)
+				h->timeout_put(t);
 			RCU_INIT_POINTER(timeout_ext->timeout, NULL);
 		}
 	}
net/netfilter/nfnetlink_cthelper.c +7 -3
index 5c622f55c9d6..97248963a7d3 100644
@@ -96,11 +96,13 @@ static int
 nfnl_cthelper_from_nlattr(struct nlattr *attr, struct nf_conn *ct)
 {
 	struct nf_conn_help *help = nfct_help(ct);
+	const struct nf_conntrack_helper *helper;
 
 	if (attr == NULL)
 		return -EINVAL;
 
-	if (help->helper->data_len == 0)
+	helper = rcu_dereference(help->helper);
+	if (!helper || helper->data_len == 0)
 		return -EINVAL;
 
 	nla_memcpy(help->data, attr, sizeof(help->data));
@@ -111,9 +113,11 @@ static int
 nfnl_cthelper_to_nlattr(struct sk_buff *skb, const struct nf_conn *ct)
 {
 	const struct nf_conn_help *help = nfct_help(ct);
+	const struct nf_conntrack_helper *helper;
 
-	if (help->helper->data_len &&
-	    nla_put(skb, CTA_HELP_INFO, help->helper->data_len, &help->data))
+	helper = rcu_dereference(help->helper);
+	if (helper && helper->data_len &&
+	    nla_put(skb, CTA_HELP_INFO, helper->data_len, &help->data))
 		goto nla_put_failure;
 
 	return 0;
net/netfilter/xt_CT.c +18 -5
index 267757b0392a..2be2f7a7b60f 100644
@@ -96,7 +96,7 @@ xt_ct_set_helper(struct nf_conn *ct, const char *helper_name,
 		return -ENOMEM;
 	}
 
-	help->helper = helper;
+	rcu_assign_pointer(help->helper, helper);
 	return 0;
 }
 
@@ -136,6 +136,21 @@ static u16 xt_ct_flags_to_dir(const struct xt_ct_target_info_v1 *info)
 	}
 }
 
+static void xt_ct_put_helper(struct nf_conn_help *help)
+{
+	struct nf_conntrack_helper *helper;
+
+	if (!help)
+		return;
+
+	/* not yet exposed to other cpus, or ruleset
+	 * already detached (post-replacement).
+	 */
+	helper = rcu_dereference_raw(help->helper);
+	if (helper)
+		nf_conntrack_helper_put(helper);
+}
+
 static int xt_ct_tg_check(const struct xt_tgchk_param *par,
 			  struct xt_ct_target_info_v1 *info)
 {
@@ -207,8 +222,7 @@ static int xt_ct_tg_check(const struct xt_tgchk_param *par,
 
 err4:
 	help = nfct_help(ct);
-	if (help)
-		nf_conntrack_helper_put(help->helper);
+	xt_ct_put_helper(help);
 err3:
 	nf_ct_tmpl_free(ct);
 err2:
@@ -270,8 +284,7 @@ static void xt_ct_tg_destroy(const struct xt_tgdtor_param *par,
 
 	if (ct) {
 		help = nfct_help(ct);
-		if (help)
-			nf_conntrack_helper_put(help->helper);
+		xt_ct_put_helper(help);
 
 		nf_ct_netns_put(par->net, par->family);
 
Signature · 1 line
2.30.2
author maintainer
#8 lore raw

PATCH 7/18 netfilter: h323: merge nat hook pointers into one

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

sparse complains about incorrect rcu usage.

Code uses the correct rcu access primitives, but the function pointers lack rcu annotations.

Collapse all of them into a single structure, then annotate the pointer.

---
 include/linux/netfilter/nf_conntrack_h323.h | 109 ++++----
 net/ipv4/netfilter/nf_nat_h323.c            |  42 ++--
 net/netfilter/nf_conntrack_h323_main.c      | 260 ++++++++------------
 3 files changed, 169 insertions(+), 242 deletions(-)
include/linux/netfilter/nf_conntrack_h323.h +56 -53
index 4561ec0fcea4..9e937f64a1ad 100644
@@ -38,60 +38,63 @@ void nf_conntrack_h245_expect(struct nf_conn *new,
 			      struct nf_conntrack_expect *this);
 void nf_conntrack_q931_expect(struct nf_conn *new,
 			      struct nf_conntrack_expect *this);
-extern int (*set_h245_addr_hook) (struct sk_buff *skb, unsigned int protoff,
-				  unsigned char **data, int dataoff,
-				  H245_TransportAddress *taddr,
-				  union nf_inet_addr *addr,
-				  __be16 port);
-extern int (*set_h225_addr_hook) (struct sk_buff *skb, unsigned int protoff,
-				  unsigned char **data, int dataoff,
-				  TransportAddress *taddr,
-				  union nf_inet_addr *addr,
-				  __be16 port);
-extern int (*set_sig_addr_hook) (struct sk_buff *skb,
-				 struct nf_conn *ct,
-				 enum ip_conntrack_info ctinfo,
-				 unsigned int protoff, unsigned char **data,
-				 TransportAddress *taddr, int count);
-extern int (*set_ras_addr_hook) (struct sk_buff *skb,
-				 struct nf_conn *ct,
-				 enum ip_conntrack_info ctinfo,
-				 unsigned int protoff, unsigned char **data,
-				 TransportAddress *taddr, int count);
-extern int (*nat_rtp_rtcp_hook) (struct sk_buff *skb,
-				 struct nf_conn *ct,
-				 enum ip_conntrack_info ctinfo,
-				 unsigned int protoff, unsigned char **data,
-				 int dataoff,
-				 H245_TransportAddress *taddr,
-				 __be16 port, __be16 rtp_port,
-				 struct nf_conntrack_expect *rtp_exp,
-				 struct nf_conntrack_expect *rtcp_exp);
-extern int (*nat_t120_hook) (struct sk_buff *skb, struct nf_conn *ct,
-			     enum ip_conntrack_info ctinfo,
-			     unsigned int protoff,
+
+struct nfct_h323_nat_hooks {
+	int (*set_h245_addr)(struct sk_buff *skb, unsigned int protoff,
 			     unsigned char **data, int dataoff,
-			     H245_TransportAddress *taddr, __be16 port,
-			     struct nf_conntrack_expect *exp);
-extern int (*nat_h245_hook) (struct sk_buff *skb, struct nf_conn *ct,
-			     enum ip_conntrack_info ctinfo,
-			     unsigned int protoff,
+			     H245_TransportAddress *taddr,
+			     union nf_inet_addr *addr, __be16 port);
+	int (*set_h225_addr)(struct sk_buff *skb, unsigned int protoff,
 			     unsigned char **data, int dataoff,
-			     TransportAddress *taddr, __be16 port,
-			     struct nf_conntrack_expect *exp);
-extern int (*nat_callforwarding_hook) (struct sk_buff *skb,
-				       struct nf_conn *ct,
-				       enum ip_conntrack_info ctinfo,
-				       unsigned int protoff,
-				       unsigned char **data, int dataoff,
-				       TransportAddress *taddr,
-				       __be16 port,
-				       struct nf_conntrack_expect *exp);
-extern int (*nat_q931_hook) (struct sk_buff *skb, struct nf_conn *ct,
-			     enum ip_conntrack_info ctinfo,
-			     unsigned int protoff,
-			     unsigned char **data, TransportAddress *taddr,
-			     int idx, __be16 port,
-			     struct nf_conntrack_expect *exp);
+			     TransportAddress *taddr,
+			     union nf_inet_addr *addr, __be16 port);
+	int (*set_sig_addr)(struct sk_buff *skb,
+			    struct nf_conn *ct,
+			    enum ip_conntrack_info ctinfo,
+			    unsigned int protoff, unsigned char **data,
+			    TransportAddress *taddr, int count);
+	int (*set_ras_addr)(struct sk_buff *skb,
+			    struct nf_conn *ct,
+			    enum ip_conntrack_info ctinfo,
+			    unsigned int protoff, unsigned char **data,
+			    TransportAddress *taddr, int count);
+	int (*nat_rtp_rtcp)(struct sk_buff *skb,
+			    struct nf_conn *ct,
+			    enum ip_conntrack_info ctinfo,
+			    unsigned int protoff,
+			    unsigned char **data, int dataoff,
+			    H245_TransportAddress *taddr,
+			    __be16 port, __be16 rtp_port,
+			    struct nf_conntrack_expect *rtp_exp,
+			    struct nf_conntrack_expect *rtcp_exp);
+	int (*nat_t120)(struct sk_buff *skb,
+			struct nf_conn *ct,
+			enum ip_conntrack_info ctinfo,
+			unsigned int protoff,
+			unsigned char **data, int dataoff,
+			H245_TransportAddress *taddr, __be16 port,
+			struct nf_conntrack_expect *exp);
+	int (*nat_h245)(struct sk_buff *skb,
+			struct nf_conn *ct,
+			enum ip_conntrack_info ctinfo,
+			unsigned int protoff,
+			unsigned char **data, int dataoff,
+			TransportAddress *taddr, __be16 port,
+			struct nf_conntrack_expect *exp);
+	int (*nat_callforwarding)(struct sk_buff *skb,
+				  struct nf_conn *ct,
+				  enum ip_conntrack_info ctinfo,
+				  unsigned int protoff,
+				  unsigned char **data, int dataoff,
+				  TransportAddress *taddr, __be16 port,
+				  struct nf_conntrack_expect *exp);
+	int (*nat_q931)(struct sk_buff *skb,
+			struct nf_conn *ct,
+			enum ip_conntrack_info ctinfo,
+			unsigned int protoff,
+			unsigned char **data, TransportAddress *taddr, int idx,
+			__be16 port, struct nf_conntrack_expect *exp);
+};
+extern const struct nfct_h323_nat_hooks __rcu *nfct_h323_nat_hook;
 
 #endif
net/ipv4/netfilter/nf_nat_h323.c +14 -28

57 lines not on this page · raw message

net/netfilter/nf_conntrack_h323_main.c +99 -161

540 lines not on this page · raw message

Signature · 1 line
2.30.2
author maintainer
#9 lore raw

PATCH 8/18 netfilter: nft_set_bitmap: Fix spelling mistake

In reply to Pablo Neira Ayuso

From: Zhang Jiaming <[email protected]>

Change 'succesful' to 'successful'.
Change 'transation' to 'transaction'.

---
 net/netfilter/nft_set_bitmap.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/netfilter/nft_set_bitmap.c +2 -2
index e7ae5914971e..96081ac8d2b4 100644
@@ -21,7 +21,7 @@ struct nft_bitmap_elem {
  * the element state in the current and the future generation.
  *
  * An element can be in three states. The generation cursor is represented using
- * the ^ character, note that this cursor shifts on every succesful transaction.
+ * the ^ character, note that this cursor shifts on every successful transaction.
  * If no transaction is going on, we observe all elements are in the following
  * state:
  *
@@ -39,7 +39,7 @@ struct nft_bitmap_elem {
  * 10 = this element is active in the current generation and it becomes inactive
  * ^    in the next one. This happens when the element is deactivated but commit
  *      path has not yet been executed yet, so removal is still pending. On
- *      transation abortion, the next generation bit is reset to go back to
+ *      transaction abortion, the next generation bit is reset to go back to
  *      restore its previous state.
  */
 struct nft_bitmap {
Signature · 1 line
2.30.2
author maintainer
#10 lore raw

PATCH 9/18 netfilter: nfnetlink: add missing __be16 cast

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Sparse flags this as suspicious, because this compares integer with a be16 with no conversion.

Its a compat check for old userspace that sends host byte order, so force a be16 cast here.

---
 net/netfilter/nfnetlink.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/netfilter/nfnetlink.c +1 -1
index 2f7c477fc9e7..c24b1240908f 100644
@@ -626,7 +626,7 @@ static void nfnetlink_rcv_skb_batch(struct sk_buff *skb, struct nlmsghdr *nlh)
 	nfgenmsg = nlmsg_data(nlh);
 	skb_pull(skb, msglen);
 	/* Work around old nft using host byte order */
-	if (nfgenmsg->res_id == NFNL_SUBSYS_NFTABLES)
+	if (nfgenmsg->res_id == (__force __be16)NFNL_SUBSYS_NFTABLES)
 		res_id = NFNL_SUBSYS_NFTABLES;
 	else
 		res_id = ntohs(nfgenmsg->res_id);
Signature · 1 line
2.30.2
author maintainer
#11 lore raw

PATCH 10/18 netfilter: x_tables: use correct integer types

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Sparse complains because __be32 and u32 are mixed without conversions. Use the correct types, no code changes.

Furthermore, xt_DSCP generates a bit truncation warning:
"cast truncates bits from constant value (ffffff03 becomes 3)"

The truncation is fine (and wanted). Add a private definition and use that instead.

objdiff shows no changes.

---
 net/netfilter/xt_DSCP.c      | 8 ++++----
 net/netfilter/xt_TCPMSS.c    | 4 ++--
 net/netfilter/xt_connlimit.c | 6 +++---
 3 files changed, 9 insertions(+), 9 deletions(-)
net/netfilter/xt_DSCP.c +4 -4
index eababc354ff1..cfa44515ab72 100644
@@ -24,6 +24,8 @@ MODULE_ALIAS("ip6t_DSCP");
 MODULE_ALIAS("ipt_TOS");
 MODULE_ALIAS("ip6t_TOS");
 
+#define XT_DSCP_ECN_MASK	3u
+
 static unsigned int
 dscp_tg(struct sk_buff *skb, const struct xt_action_param *par)
 {
@@ -34,8 +36,7 @@ dscp_tg(struct sk_buff *skb, const struct xt_action_param *par)
 		if (skb_ensure_writable(skb, sizeof(struct iphdr)))
 			return NF_DROP;
 
-		ipv4_change_dsfield(ip_hdr(skb),
-				    (__force __u8)(~XT_DSCP_MASK),
+		ipv4_change_dsfield(ip_hdr(skb), XT_DSCP_ECN_MASK,
 				    dinfo->dscp << XT_DSCP_SHIFT);
 
 	}
@@ -52,8 +53,7 @@ dscp_tg6(struct sk_buff *skb, const struct xt_action_param *par)
 		if (skb_ensure_writable(skb, sizeof(struct ipv6hdr)))
 			return NF_DROP;
 
-		ipv6_change_dsfield(ipv6_hdr(skb),
-				    (__force __u8)(~XT_DSCP_MASK),
+		ipv6_change_dsfield(ipv6_hdr(skb), XT_DSCP_ECN_MASK,
 				    dinfo->dscp << XT_DSCP_SHIFT);
 	}
 	return XT_CONTINUE;
net/netfilter/xt_TCPMSS.c +2 -2
index 122db9fbb9f4..116a885adb3c 100644
@@ -239,8 +239,8 @@ tcpmss_tg6(struct sk_buff *skb, const struct xt_action_param *par)
 		oldlen = ipv6h->payload_len;
 		newlen = htons(ntohs(oldlen) + ret);
 		if (skb->ip_summed == CHECKSUM_COMPLETE)
-			skb->csum = csum_add(csum_sub(skb->csum, oldlen),
-					     newlen);
+			skb->csum = csum_add(csum_sub(skb->csum, (__force __wsum)oldlen),
+					     (__force __wsum)newlen);
 		ipv6h->payload_len = newlen;
 	}
 	return XT_CONTINUE;
net/netfilter/xt_connlimit.c +3 -3
index 46fcac75f726..5d04ef80a61d 100644
@@ -62,10 +62,10 @@ connlimit_mt(const struct sk_buff *skb, struct xt_action_param *par)
 		key[4] = zone->id;
 	} else {
 		const struct iphdr *iph = ip_hdr(skb);
-		key[0] = (info->flags & XT_CONNLIMIT_DADDR) ?
-			  iph->daddr : iph->saddr;
 
-		key[0] &= info->mask.ip;
+		key[0] = (info->flags & XT_CONNLIMIT_DADDR) ?
+			 (__force __u32)iph->daddr : (__force __u32)iph->saddr;
+		key[0] &= (__force __u32)info->mask.ip;
 		key[1] = zone->id;
 	}
 
Signature · 1 line
2.30.2
author maintainer
#12 lore raw

PATCH 11/18 netfilter: nf_tables: use the correct get/put helpers

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Switch to be16/32 and u16/32 respectively. No code changes here, the functions do the same thing, this is just for sparse checkers' sake.

objdiff shows no changes.

---
 net/netfilter/nft_byteorder.c | 3 ++-
 net/netfilter/nft_osf.c       | 2 +-
 net/netfilter/nft_socket.c    | 8 ++++----
 net/netfilter/nft_xfrm.c      | 8 ++++----
 4 files changed, 11 insertions(+), 10 deletions(-)
net/netfilter/nft_byteorder.c +2 -1
index d77609144b26..f952a80275a8 100644
@@ -44,7 +44,8 @@ void nft_byteorder_eval(const struct nft_expr *expr,
 		case NFT_BYTEORDER_NTOH:
 			for (i = 0; i < priv->len / 8; i++) {
 				src64 = nft_reg_load64(&src[i]);
-				nft_reg_store64(&dst[i], be64_to_cpu(src64));
+				nft_reg_store64(&dst[i],
+						be64_to_cpu((__force __be64)src64));
 			}
 			break;
 		case NFT_BYTEORDER_HTON:
net/netfilter/nft_osf.c +1 -1
index 5eed18f90b02..0053a697c931 100644
@@ -99,7 +99,7 @@ static int nft_osf_dump(struct sk_buff *skb, const struct nft_expr *expr)
 	if (nla_put_u8(skb, NFTA_OSF_TTL, priv->ttl))
 		goto nla_put_failure;
 
-	if (nla_put_be32(skb, NFTA_OSF_FLAGS, ntohl(priv->flags)))
+	if (nla_put_u32(skb, NFTA_OSF_FLAGS, ntohl((__force __be32)priv->flags)))
 		goto nla_put_failure;
 
 	if (nft_dump_register(skb, NFTA_OSF_DREG, priv->dreg))
net/netfilter/nft_socket.c +4 -4
index 05ae5a338b6f..a7de29137618 100644
@@ -163,7 +163,7 @@ static int nft_socket_init(const struct nft_ctx *ctx,
 		return -EOPNOTSUPP;
 	}
 
-	priv->key = ntohl(nla_get_u32(tb[NFTA_SOCKET_KEY]));
+	priv->key = ntohl(nla_get_be32(tb[NFTA_SOCKET_KEY]));
 	switch(priv->key) {
 	case NFT_SOCKET_TRANSPARENT:
 	case NFT_SOCKET_WILDCARD:
@@ -179,7 +179,7 @@ static int nft_socket_init(const struct nft_ctx *ctx,
 		if (!tb[NFTA_SOCKET_LEVEL])
 			return -EINVAL;
 
-		level = ntohl(nla_get_u32(tb[NFTA_SOCKET_LEVEL]));
+		level = ntohl(nla_get_be32(tb[NFTA_SOCKET_LEVEL]));
 		if (level > 255)
 			return -EOPNOTSUPP;
 
@@ -202,12 +202,12 @@ static int nft_socket_dump(struct sk_buff *skb,
 {
 	const struct nft_socket *priv = nft_expr_priv(expr);
 
-	if (nla_put_u32(skb, NFTA_SOCKET_KEY, htonl(priv->key)))
+	if (nla_put_be32(skb, NFTA_SOCKET_KEY, htonl(priv->key)))
 		return -1;
 	if (nft_dump_register(skb, NFTA_SOCKET_DREG, priv->dreg))
 		return -1;
 	if (priv->key == NFT_SOCKET_CGROUPV2 &&
-	    nla_put_u32(skb, NFTA_SOCKET_LEVEL, htonl(priv->level)))
+	    nla_put_be32(skb, NFTA_SOCKET_LEVEL, htonl(priv->level)))
 		return -1;
 	return 0;
 }
net/netfilter/nft_xfrm.c +4 -4
index becb88fa4e9b..1c5343c936a8 100644
@@ -51,7 +51,7 @@ static int nft_xfrm_get_init(const struct nft_ctx *ctx,
 		return -EOPNOTSUPP;
 	}
 
-	priv->key = ntohl(nla_get_u32(tb[NFTA_XFRM_KEY]));
+	priv->key = ntohl(nla_get_be32(tb[NFTA_XFRM_KEY]));
 	switch (priv->key) {
 	case NFT_XFRM_KEY_REQID:
 	case NFT_XFRM_KEY_SPI:
@@ -134,13 +134,13 @@ static void nft_xfrm_state_get_key(const struct nft_xfrm *priv,
 		WARN_ON_ONCE(1);
 		break;
 	case NFT_XFRM_KEY_DADDR_IP4:
-		*dest = state->id.daddr.a4;
+		*dest = (__force __u32)state->id.daddr.a4;
 		return;
 	case NFT_XFRM_KEY_DADDR_IP6:
 		memcpy(dest, &state->id.daddr.in6, sizeof(struct in6_addr));
 		return;
 	case NFT_XFRM_KEY_SADDR_IP4:
-		*dest = state->props.saddr.a4;
+		*dest = (__force __u32)state->props.saddr.a4;
 		return;
 	case NFT_XFRM_KEY_SADDR_IP6:
 		memcpy(dest, &state->props.saddr.in6, sizeof(struct in6_addr));
@@ -149,7 +149,7 @@ static void nft_xfrm_state_get_key(const struct nft_xfrm *priv,
 		*dest = state->props.reqid;
 		return;
 	case NFT_XFRM_KEY_SPI:
-		*dest = state->id.spi;
+		*dest = (__force __u32)state->id.spi;
 		return;
 	}
 
Signature · 1 line
2.30.2
author maintainer
#13 lore raw

PATCH 12/18 netfilter: nf_tables: add and use BE register load-store helpers

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Same as the existing ones, no conversions. This is just for sparse sake only so that we no longer mix be16/u16 and be32/u32 types.

Alternative is to add __force __beX in various places, but this seems nicer.

objdiff shows no changes.

---
 include/net/netfilter/nf_tables.h      | 15 +++++++++++++++
 net/bridge/netfilter/nft_meta_bridge.c |  2 +-
 net/netfilter/nft_tproxy.c             |  6 +++---
 3 files changed, 19 insertions(+), 4 deletions(-)
include/net/netfilter/nf_tables.h +15 -0
index 5c4e5a96a984..dd6ad0b67d7d 100644
@@ -157,11 +157,26 @@ static inline void nft_reg_store16(u32 *dreg, u16 val)
 	*(u16 *)dreg = val;
 }
 
+static inline void nft_reg_store_be16(u32 *dreg, __be16 val)
+{
+	nft_reg_store16(dreg, (__force __u16)val);
+}
+
 static inline u16 nft_reg_load16(const u32 *sreg)
 {
 	return *(u16 *)sreg;
 }
 
+static inline __be16 nft_reg_load_be16(const u32 *sreg)
+{
+	return (__force __be16)nft_reg_load16(sreg);
+}
+
+static inline __be32 nft_reg_load_be32(const u32 *sreg)
+{
+	return *(__force __be32 *)sreg;
+}
+
 static inline void nft_reg_store64(u32 *dreg, u64 val)
 {
 	put_unaligned(val, (u64 *)dreg);
net/bridge/netfilter/nft_meta_bridge.c +1 -1
index 8c3eaba87ad2..c3ecd77e25cb 100644
@@ -53,7 +53,7 @@ static void nft_meta_bridge_get_eval(const struct nft_expr *expr,
 			goto err;
 
 		br_vlan_get_proto(br_dev, &p_proto);
-		nft_reg_store16(dest, htons(p_proto));
+		nft_reg_store_be16(dest, htons(p_proto));
 		return;
 	}
 	default:
net/netfilter/nft_tproxy.c +3 -3
index 801f013971df..68b2eed742df 100644
@@ -52,11 +52,11 @@ static void nft_tproxy_eval_v4(const struct nft_expr *expr,
 				   skb->dev, NF_TPROXY_LOOKUP_ESTABLISHED);
 
 	if (priv->sreg_addr)
-		taddr = regs->data[priv->sreg_addr];
+		taddr = nft_reg_load_be32(&regs->data[priv->sreg_addr]);
 	taddr = nf_tproxy_laddr4(skb, taddr, iph->daddr);
 
 	if (priv->sreg_port)
-		tport = nft_reg_load16(&regs->data[priv->sreg_port]);
+		tport = nft_reg_load_be16(&regs->data[priv->sreg_port]);
 	if (!tport)
 		tport = hp->dest;
 
@@ -124,7 +124,7 @@ static void nft_tproxy_eval_v6(const struct nft_expr *expr,
 	taddr = *nf_tproxy_laddr6(skb, &taddr, &iph->daddr);
 
 	if (priv->sreg_port)
-		tport = nft_reg_load16(&regs->data[priv->sreg_port]);
+		tport = nft_reg_load_be16(&regs->data[priv->sreg_port]);
 	if (!tport)
 		tport = hp->dest;
 
Signature · 1 line
2.30.2
author maintainer
#14 lore raw

PATCH 13/18 netfilter: nf_tables: use correct integer types

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

Sparse tool complains about mixing of different endianess types, so use the correct ones.

Add type casts where needed.

objdiff shows no changes except in nft_tunnel (type is changed).

---
 net/netfilter/nft_cmp.c    |  6 +++---
 net/netfilter/nft_ct.c     |  4 ++--
 net/netfilter/nft_exthdr.c | 10 +++++-----
 net/netfilter/nft_tunnel.c |  3 ++-
 4 files changed, 12 insertions(+), 11 deletions(-)
net/netfilter/nft_cmp.c +3 -3
index 6528f76ca29e..bec22584395f 100644
@@ -125,13 +125,13 @@ static void nft_payload_n2h(union nft_cmp_offload_data *data,
 {
 	switch (len) {
 	case 2:
-		data->val16 = ntohs(*((u16 *)val));
+		data->val16 = ntohs(*((__be16 *)val));
 		break;
 	case 4:
-		data->val32 = ntohl(*((u32 *)val));
+		data->val32 = ntohl(*((__be32 *)val));
 		break;
 	case 8:
-		data->val64 = be64_to_cpu(*((u64 *)val));
+		data->val64 = be64_to_cpu(*((__be64 *)val));
 		break;
 	default:
 		WARN_ON_ONCE(1);
net/netfilter/nft_ct.c +2 -2
index d8e1614918a1..b04995c3e17f 100644
@@ -204,12 +204,12 @@ static void nft_ct_get_eval(const struct nft_expr *expr,
 	case NFT_CT_SRC_IP:
 		if (nf_ct_l3num(ct) != NFPROTO_IPV4)
 			goto err;
-		*dest = tuple->src.u3.ip;
+		*dest = (__force __u32)tuple->src.u3.ip;
 		return;
 	case NFT_CT_DST_IP:
 		if (nf_ct_l3num(ct) != NFPROTO_IPV4)
 			goto err;
-		*dest = tuple->dst.u3.ip;
+		*dest = (__force __u32)tuple->dst.u3.ip;
 		return;
 	case NFT_CT_SRC_IP6:
 		if (nf_ct_l3num(ct) != NFPROTO_IPV6)
net/netfilter/nft_exthdr.c +5 -5
index 22c3e05b52db..a67ea9c3ae57 100644
@@ -266,7 +266,7 @@ static void nft_exthdr_tcp_set_eval(const struct nft_expr *expr,
 
 		switch (priv->len) {
 		case 2:
-			old.v16 = get_unaligned((u16 *)(opt + offset));
+			old.v16 = (__force __be16)get_unaligned((u16 *)(opt + offset));
 			new.v16 = (__force __be16)nft_reg_load16(
 				&regs->data[priv->sreg]);
 
@@ -281,18 +281,18 @@ static void nft_exthdr_tcp_set_eval(const struct nft_expr *expr,
 			if (old.v16 == new.v16)
 				return;
 
-			put_unaligned(new.v16, (u16*)(opt + offset));
+			put_unaligned(new.v16, (__be16*)(opt + offset));
 			inet_proto_csum_replace2(&tcph->check, pkt->skb,
 						 old.v16, new.v16, false);
 			break;
 		case 4:
-			new.v32 = regs->data[priv->sreg];
-			old.v32 = get_unaligned((u32 *)(opt + offset));
+			new.v32 = nft_reg_load_be32(&regs->data[priv->sreg]);
+			old.v32 = (__force __be32)get_unaligned((u32 *)(opt + offset));
 
 			if (old.v32 == new.v32)
 				return;
 
-			put_unaligned(new.v32, (u32*)(opt + offset));
+			put_unaligned(new.v32, (__be32*)(opt + offset));
 			inet_proto_csum_replace4(&tcph->check, pkt->skb,
 						 old.v32, new.v32, false);
 			break;
net/netfilter/nft_tunnel.c +2 -1
index d0f9b1d51b0e..5edaaded706d 100644
@@ -383,8 +383,9 @@ static int nft_tunnel_obj_opts_init(const struct nft_ctx *ctx,
 				    struct ip_tunnel_info *info,
 				    struct nft_tunnel_opts *opts)
 {
-	int err, rem, type = 0;
 	struct nlattr *nla;
+	__be16 type = 0;
+	int err, rem;
 
 	err = nla_validate_nested_deprecated(attr, NFTA_TUNNEL_KEY_OPTS_MAX,
 					     nft_tunnel_opts_policy, NULL);
Signature · 1 line
2.30.2
author maintainer
#15 lore raw

PATCH 14/18 netfilter: nf_tables: move nft_cmp_fast_mask to where its used

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

... and cast result to u32 so sparse won't complain anymore.

---
 include/net/netfilter/nf_tables_core.h | 10 ----------
 net/netfilter/nft_cmp.c                | 12 ++++++++++++
 2 files changed, 12 insertions(+), 10 deletions(-)
include/net/netfilter/nf_tables_core.h +0 -10
index 0ea7c55cea4d..1223af68cd9a 100644
@@ -56,16 +56,6 @@ struct nft_immediate_expr {
 	u8			dlen;
 };
 
-/* Calculate the mask for the nft_cmp_fast expression. On big endian the
- * mask needs to include the *upper* bytes when interpreting that data as
- * something smaller than the full u32, therefore a cpu_to_le32 is done.
- */
-static inline u32 nft_cmp_fast_mask(unsigned int len)
-{
-	return cpu_to_le32(~0U >> (sizeof_field(struct nft_cmp_fast_expr,
-						data) * BITS_PER_BYTE - len));
-}
-
 extern const struct nft_expr_ops nft_cmp_fast_ops;
 extern const struct nft_expr_ops nft_cmp16_fast_ops;
 
net/netfilter/nft_cmp.c +12 -0
index bec22584395f..777f09e4dc60 100644
@@ -197,6 +197,18 @@ static const struct nft_expr_ops nft_cmp_ops = {
 	.offload	= nft_cmp_offload,
 };
 
+/* Calculate the mask for the nft_cmp_fast expression. On big endian the
+ * mask needs to include the *upper* bytes when interpreting that data as
+ * something smaller than the full u32, therefore a cpu_to_le32 is done.
+ */
+static u32 nft_cmp_fast_mask(unsigned int len)
+{
+	__le32 mask = cpu_to_le32(~0U >> (sizeof_field(struct nft_cmp_fast_expr,
+					  data) * BITS_PER_BYTE - len));
+
+	return (__force u32)mask;
+}
+
 static int nft_cmp_fast_init(const struct nft_ctx *ctx,
 			     const struct nft_expr *expr,
 			     const struct nlattr * const tb[])
Signature · 1 line
2.30.2
author maintainer
#16 lore raw

PATCH 15/18 netfilter: nf_nat: in nf_nat_initialized(), use const struct nf_conn *

In reply to Pablo Neira Ayuso

From: James Yonan <[email protected]>

nf_nat_initialized() doesn't modify passed struct nf_conn, so declare as const.

This is helpful for code readability and makes it possible to call nf_nat_initialized() with a const struct nf_conn *.

---
 include/net/netfilter/nf_nat.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/net/netfilter/nf_nat.h +1 -1
index 987111ae5240..e9eb01e99d2f 100644
@@ -104,7 +104,7 @@ unsigned int
 nf_nat_inet_fn(void *priv, struct sk_buff *skb,
 	       const struct nf_hook_state *state);
 
-static inline int nf_nat_initialized(struct nf_conn *ct,
+static inline int nf_nat_initialized(const struct nf_conn *ct,
 				     enum nf_nat_manip_type manip)
 {
 	if (manip == NF_NAT_MANIP_SRC)
Signature · 1 line
2.30.2
author maintainer
#17 lore raw

PATCH 16/18 netfilter: ipvs: Use the bitmap API to allocate bitmaps

In reply to Pablo Neira Ayuso

From: Christophe JAILLET <[email protected]>

Use bitmap_zalloc()/bitmap_free() instead of hand-writing them.

It is less verbose and it improves the semantic.

---
 net/netfilter/ipvs/ip_vs_mh.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
net/netfilter/ipvs/ip_vs_mh.c +2 -3
index da0280cec506..e3d7f5c879ce 100644
@@ -174,8 +174,7 @@ static int ip_vs_mh_populate(struct ip_vs_mh_state *s,
 		return 0;
 	}
 
-	table = kcalloc(BITS_TO_LONGS(IP_VS_MH_TAB_SIZE),
-			sizeof(unsigned long), GFP_KERNEL);
+	table = bitmap_zalloc(IP_VS_MH_TAB_SIZE, GFP_KERNEL);
 	if (!table)
 		return -ENOMEM;
 
@@ -227,7 +226,7 @@ static int ip_vs_mh_populate(struct ip_vs_mh_state *s,
 	}
 
 out:
-	kfree(table);
+	bitmap_free(table);
 	return 0;
 }
 
Signature · 1 line
2.30.2
author maintainer
#18 lore raw

PATCH 17/18 netfilter: flowtable: prefer refcount_inc

In reply to Pablo Neira Ayuso

From: Florian Westphal <[email protected]>

With refcount_inc_not_zero, we'd also need a smp_rmb or similar, followed by a test of the CONFIRMED bit.

However, the ct pointer is taken from skb->_nfct, its refcount must not be 0 (else, we'd already have a use-after-free bug).

Use refcount_inc() instead to clarify the ct refcount is expected to be at least 1.

---
 net/netfilter/nf_flow_table_core.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)
net/netfilter/nf_flow_table_core.c +3 -8
index 51c2b1570838..765ac779bfc8 100644
@@ -53,14 +53,14 @@ struct flow_offload *flow_offload_alloc(struct nf_conn *ct)
 {
 	struct flow_offload *flow;
 
-	if (unlikely(nf_ct_is_dying(ct) ||
-	    !refcount_inc_not_zero(&ct->ct_general.use)))
+	if (unlikely(nf_ct_is_dying(ct)))
 		return NULL;
 
 	flow = kzalloc(sizeof(*flow), GFP_ATOMIC);
 	if (!flow)
-		goto err_ct_refcnt;
+		return NULL;
 
+	refcount_inc(&ct->ct_general.use);
 	flow->ct = ct;
 
 	flow_offload_fill_dir(flow, FLOW_OFFLOAD_DIR_ORIGINAL);
@@ -72,11 +72,6 @@ struct flow_offload *flow_offload_alloc(struct nf_conn *ct)
 		__set_bit(NF_FLOW_DNAT, &flow->flags);
 
 	return flow;
-
-err_ct_refcnt:
-	nf_ct_put(ct);
-
-	return NULL;
 }
 EXPORT_SYMBOL_GPL(flow_offload_alloc);
 
Signature · 1 line
2.30.2
author maintainer
#19 lore raw

PATCH 18/18 netfilter: xt_TPROXY: remove pr_debug invocations

In reply to Pablo Neira Ayuso

From: Justin Stitt <[email protected]>

pr_debug calls are no longer needed in this file.

Pablo suggested "a patch to remove these pr_debug calls". This patch has some other beneficial collateral as it also silences multiple Clang -Wformat warnings that were present in the pr_debug calls.

diff from v1 -> v2:
* converted if statement one-liner style
* x == NULL is now !x

---
 net/netfilter/xt_TPROXY.c | 25 ++-----------------------
 1 file changed, 2 insertions(+), 23 deletions(-)
net/netfilter/xt_TPROXY.c +2 -23
index 459d0696c91a..e4bea1d346cf 100644
@@ -74,18 +74,10 @@ tproxy_tg4(struct net *net, struct sk_buff *skb, __be32 laddr, __be16 lport,
 		/* This should be in a separate target, but we don't do multiple
 		   targets on the same rule yet */
 		skb->mark = (skb->mark & ~mark_mask) ^ mark_value;
-
-		pr_debug("redirecting: proto %hhu %pI4:%hu -> %pI4:%hu, mark: %x\n",
-			 iph->protocol, &iph->daddr, ntohs(hp->dest),
-			 &laddr, ntohs(lport), skb->mark);
-
 		nf_tproxy_assign_sock(skb, sk);
 		return NF_ACCEPT;
 	}
 
-	pr_debug("no socket, dropping: proto %hhu %pI4:%hu -> %pI4:%hu, mark: %x\n",
-		 iph->protocol, &iph->saddr, ntohs(hp->source),
-		 &iph->daddr, ntohs(hp->dest), skb->mark);
 	return NF_DROP;
 }
 
@@ -122,16 +114,12 @@ tproxy_tg6_v1(struct sk_buff *skb, const struct xt_action_param *par)
 	int tproto;
 
 	tproto = ipv6_find_hdr(skb, &thoff, -1, NULL, NULL);
-	if (tproto < 0) {
-		pr_debug("unable to find transport header in IPv6 packet, dropping\n");
+	if (tproto < 0)
 		return NF_DROP;
-	}
 
 	hp = skb_header_pointer(skb, thoff, sizeof(_hdr), &_hdr);
-	if (hp == NULL) {
-		pr_debug("unable to grab transport header contents in IPv6 packet, dropping\n");
+	if (!hp)
 		return NF_DROP;
-	}
 
 	/* check if there's an ongoing connection on the packet
 	 * addresses, this happens if the redirect already happened
@@ -168,19 +156,10 @@ tproxy_tg6_v1(struct sk_buff *skb, const struct xt_action_param *par)
 		/* This should be in a separate target, but we don't do multiple
 		   targets on the same rule yet */
 		skb->mark = (skb->mark & ~tgi->mark_mask) ^ tgi->mark_value;
-
-		pr_debug("redirecting: proto %hhu %pI6:%hu -> %pI6:%hu, mark: %x\n",
-			 tproto, &iph->saddr, ntohs(hp->source),
-			 laddr, ntohs(lport), skb->mark);
-
 		nf_tproxy_assign_sock(skb, sk);
 		return NF_ACCEPT;
 	}
 
-	pr_debug("no socket, dropping: proto %hhu %pI6:%hu -> %pI6:%hu, mark: %x\n",
-		 tproto, &iph->saddr, ntohs(hp->source),
-		 &iph->daddr, ntohs(hp->dest), skb->mark);
-
 	return NF_DROP;
 }
 
Signature · 1 line
2.30.2