Skip to contentSkip to conversation

Passive OS fingerprint xtables match.

PATCH Evgeniy Polyakov 5 participants 11 messages Mar 10, 2009 – Mar 16, 2009

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

Hi.

Passive OS fingerprinting netfilter module allows to passively detect remote OS and perform various netfilter actions based on that knowledge. This module compares some data (WS, MSS, options and it's order, ttl, df and others) from packets with SYN bit set with dynamically loaded OS fingerprints.

Fingerprint matching rules can be downloaded from OpenBSD source tree and loaded via netlink connector into the kernel via special util found in archive. It will also listen for events about matching packets.

Archive also contains library file (also attached), which was shipped with iptables extensions some time ago (at least when ipt_osf existed in patch-o-matic).

This release implements suggestions found during the code review like codying style, structure split and tighter packing, bool and %pi4 usage and similar changes.

Fingerprints can be downloaded from
http://www.openbsd.org/cgi-bin/cvsweb/src/etc/pf.os

Example usage:

# modrpobe xt_osf
# ./ucon_osf -f ./pf.os
^C Daemon will listen for incoming match events 
-d switch removes fingerprints
# iptables -I INPUT -j ACCEPT -p tcp -m osf --genre Linux --log 0 --ttl 2 --connector

You will find something like this in the syslog:
ipt_osf: Windows [2000:SP3:Windows XP Pro SP1, 2000 SP3]: 11.22.33.55:4024 -> 11.22.33.44:139

Please consider for inclusion.
Thank you.

Passive OS fingerprint homepage (archives, examples):
http://www.ioremap.net/projects/osf

include/linux/connector.h +2 -0
index 34f2789..da6595e 100644
@@ -39,6 +39,8 @@
 #define CN_IDX_V86D			0x4
 #define CN_VAL_V86D_UVESAFB		0x1
 #define CN_IDX_BB			0x5	/* BlackBoard, from the TSP GPL sampling framework */
+#define CN_IDX_OSF			0x6	/* Passive OS fingerprint iptables module */
+#define CN_VAL_OSF			0x0
 
 #define CN_NETLINK_USERS		6
 
include/linux/netfilter/xt_osf.h new file +110 -0

110 lines not on this page · raw message

net/netfilter/Kconfig +13 -0
index c2bac9c..34aafec 100644
@@ -853,6 +853,19 @@ config NETFILTER_XT_MATCH_U32
 
 	  Details and examples are in the kernel module source.
 
+config NETFILTER_XT_MATCH_OSF
+	tristate 'Passive OS fingerprint match support'
+	depends on NETFILTER_ADVANCED
+	help
+	  Passive OS fingerprint matching module. Allows to passively match
+	  remote operation system analyzing incoming TCP packets with SYN
+	  bit set.
+
+	  You should download and install rule loading software from
+	  http://www.ioremap.net/projects/osf
+
+	  To compile it as a module, choose M here.  If unsure, say N.
+
 endif # NETFILTER_XTABLES
 
 endmenu
net/netfilter/Makefile +1 -0
index da3d909..ec3a6e0 100644
@@ -89,6 +89,7 @@ obj-$(CONFIG_NETFILTER_XT_MATCH_STRING) += xt_string.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_TCPMSS) += xt_tcpmss.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_TIME) += xt_time.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_U32) += xt_u32.o
+obj-$(CONFIG_NETFILTER_XT_MATCH_OSF) += xt_osf.o
 
 # IPVS
 obj-$(CONFIG_IP_VS) += ipvs/
net/netfilter/xt_osf.c new file +475 -0

475 lines not on this page · raw message

Signature · 1 line
	Evgeniy Polyakov

/*
* Copyright (c) 2003+ Evgeniy Polyakov <[email protected]>
*
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */

/*
* iptables interface for OS fingerprint matching module. */

#include <stdio.h>
#include <netdb.h>
#include <string.h>
#include <stdlib.h>
#include <getopt.h>
#include <ctype.h>

#include <linux/types.h>

#include <xtables.h>

#include <netinet/ip.h>
#include <netinet/tcp.h>

#include <linux/netfilter/xt_osf.h>

static void osf_help(void)
{
printf("OS fingerprint match options:\n"
"--genre [!] string Match a OS genre by passive fingerprinting.\n"
"--ttl Use some TTL check extensions to determine OS:\n"
" 0 true ip and fingerprint TTL comparison. Works for LAN.\n"
" 1 check if ip TTL is less than fingerprint one. Works for global addresses.\n"
" 2 do not compare TTL at all. Allows to detect NMAP, but can produce false results.\n"
"--log level Log determined genres into dmesg even if they do not match desired one:\n"
" 0 log all matched or unknown signatures.\n"
" 1 log only first one.\n"
" 2 log all known matched signatures.\n"
"--connector Log through kernel connector.\n\n"
);
}

static const struct option osf_opts[] = {
	{ .name = "genre",	.has_arg = true, .val = '1' },
	{ .name = "ttl",	.has_arg = true, .val = '2' },
	{ .name = "log",	.has_arg = true, .val = '3' },
	{ .name = "connector",	.has_arg = false, .val = '5' },
	{ .name = NULL }
};

static void osf_init(struct xt_entry_match *m)
{
}

static void osf_parse_string(const unsigned char *s, struct ipt_osf_info *info)

{
	if (strlen(s) < MAXGENRELEN) 
		strcpy(info->genre, s);
	else 
		exit_error(PARAMETER_PROBLEM, "Genre string too long `%s' [%d], max=%d", 
				s, strlen(s), MAXGENRELEN);
}

static int osf_parse(int c, char **argv, int invert, unsigned int *flags,
const void *entry,
struct xt_entry_match **match)

{
	struct ipt_osf_info *info = (struct ipt_osf_info *)(*match)->data;
	
	switch(c) 
	{
		case '1': /* --genre */
			if (*flags & IPT_OSF_GENRE)
				exit_error(PARAMETER_PROBLEM, "Can't specify multiple genre parameter");
			check_inverse(optarg, &invert, &optind, 0);
			osf_parse_string(argv[optind-1], info);
			if (invert)
				info->flags |= IPT_OSF_INVERT;
			info->len=strlen(info->genre);
			*flags |= IPT_OSF_GENRE;
			break;
		case '2': /* --ttl */
			if (*flags & IPT_OSF_TTL)
				exit_error(PARAMETER_PROBLEM, "Can't specify multiple ttl parameter");
			*flags |= IPT_OSF_TTL;
			info->flags |= IPT_OSF_TTL;
			info->ttl = atoi(argv[optind-1]);
			break;
		case '3': /* --log */
			if (*flags & IPT_OSF_LOG)
				exit_error(PARAMETER_PROBLEM, "Can't specify multiple log parameter");
			*flags |= IPT_OSF_LOG;
			info->loglevel = atoi(argv[optind-1]);
			info->flags |= IPT_OSF_LOG;
			break;
		case '5': /* --connector */
			if (*flags & IPT_OSF_CONNECTOR)
				exit_error(PARAMETER_PROBLEM, "Can't specify multiple connector parameter");
			*flags |= IPT_OSF_CONNECTOR;
			info->flags |= IPT_OSF_CONNECTOR;
			break;
		default:
			return 0;
	}

	return 1;
}

static void osf_final_check(unsigned int flags)

{
	if (!flags)
		exit_error(PARAMETER_PROBLEM, "OS fingerprint match: You must specify `--genre'");
}

static void osf_print(const void *ip, const struct xt_entry_match *match, int numeric)

{
	const struct ipt_osf_info *info = (const struct ipt_osf_info*) match->data;

	printf("OS fingerprint match %s%s ", (info->flags & IPT_OSF_INVERT) ? "! " : "", info->genre);
}

static void osf_save(const void *ip, const struct xt_entry_match *match)

{
	const struct ipt_osf_info *info = (const struct ipt_osf_info*) match->data;

	printf("--genre %s%s ", (info->flags & IPT_OSF_INVERT) ? "! ": "", info->genre);
}
static struct xtables_match osf_match = {
	.name		= "osf",
	.version	= XTABLES_VERSION,
	.size		= XT_ALIGN(sizeof(struct ipt_osf_info)),
	.userspacesize	= XT_ALIGN(sizeof(struct ipt_osf_info)),
	.help		= osf_help,
	.init		= osf_init,
	.parse		= osf_parse,
	.print		= osf_print,
	.final_check	= osf_final_check,
	.save		= osf_save,
	.extra_opts	= osf_opts
};

void _init(void)

{
	xtables_register_match(&osf_match);
}
#1 lore raw
Quoted 10 lines · Evgeniy Polyakov

Passive OS fingerprinting netfilter module allows to passively detect remote OS and perform various netfilter actions based on that knowledge. This module compares some data (WS, MSS, options and it's order, ttl, df and others) from packets with SYN bit set with dynamically loaded OS fingerprints.

Fingerprint matching rules can be downloaded from OpenBSD source tree and loaded via netlink connector into the kernel via special util found in archive. It will also listen for events about matching packets.

Quoted 7 lines

Archive also contains library file (also attached), which was shipped with iptables extensions some time ago (at least when ipt_osf existed in patch-o-matic).

This release implements suggestions found during the code review like codying style, structure split and tighter packing, bool and %pi4 usage and similar changes.

Not the latest version, it misses the following fix from the parallel tree.

Fixed TCP header copy to the userspace when given option is enabled.

net/netfilter/xt_osf.c +4 -2
index 3114bbd..e619f09 100644
@@ -83,7 +83,7 @@ static void ipt_osf_send_connector(struct ipt_osf_user_finger *f,
 	struct ipt_osf_message *msg = &per_cpu(ipt_osf_mbuf, smp_processor_id());
 	struct ipt_osf_nlmsg *data = &msg->nlmsg;
 	struct iphdr *iph = ip_hdr(skb);
-	struct tcphdr *tcph = tcp_hdr(skb);
+	struct tcphdr *tcp;
 
 	memcpy(&msg->cmsg.id, &cn_osf_id, sizeof(struct cn_msg));
 	msg->cmsg.seq = osf_seq++;
@@ -92,7 +92,9 @@ static void ipt_osf_send_connector(struct ipt_osf_user_finger *f,
 
 	memcpy(&data->f, f, sizeof(struct ipt_osf_user_finger));
 	memcpy(&data->ip, iph, sizeof(struct iphdr));
-	memcpy(&data->tcp, tcph, sizeof(struct tcphdr));
+	tcp = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(struct tcphdr), &data->tcp);
+	if (tcp)
+		memcpy(&data->tcp, tcp, sizeof(struct tcphdr));
 
 	cn_netlink_send(&msg->cmsg, CN_IDX_OSF, GFP_ATOMIC);
 }
Signature · 1 line
	Evgeniy Polyakov
#2 lore raw
In reply to Evgeniy Polyakov
Quoted 2 lines · Evgeniy Polyakov

[xt_osf]

This starts to get its shape. I'm in favor of it.

include/linux/netfilter/xt_osf.h line 3

Quoted 6 lines
include/linux/netfilter/xt_osf.h new file +3 -0
@@ -0,0 +1,110 @@
+/*
+ * xt_osf.h
+ *

Common-sense policy wants that filenames like these not be part of the file.

Quoted 28 lines
include/linux/netfilter/xt_osf.h new file +28 -0
+#ifndef _IPT_OSF_H
+#define _IPT_OSF_H
+
+#define MAXGENRELEN		32
+#define MAXDETLEN		64
+
+#define IPT_OSF_GENRE		(1<<0)
+#define	IPT_OSF_TTL		(1<<1)
+#define IPT_OSF_LOG		(1<<2)
+#define IPT_OSF_UNUSED		(1<<3)
+#define IPT_OSF_CONNECTOR	(1<<4)
+#define IPT_OSF_INVERT		(1<<5)
+
+#define IPT_OSF_LOGLEVEL_ALL	0
+#define IPT_OSF_LOGLEVEL_FIRST	1
+#define IPT_OSF_LOGLEVEL_ALL_KNOWN	2
+
+#define IPT_OSF_TTL_TRUE	0	/* True ip and fingerprint TTL comparison */
+#define IPT_OSF_TTL_LESS	1	/* Check if ip TTL is less than fingerprint one */
+#define IPT_OSF_TTL_NOCHECK	2	/* Do not compare ip and fingerprint TTL at all */
+
+struct ipt_osf_info {
+	char			genre[MAXGENRELEN];
+	__u32			len;
+	__u32			flags;
+	__u32			loglevel;
+	__u32			ttl;
+};
Quoted 1 line

[more struct ipt_osf...]

Since the module's name is xt_osf now, it would make sense to follow this in the struct names too.

net/netfilter/Kconfig config NETFILTER_XT_MATCH_U32 · line 857

Quoted 10 lines
net/netfilter/Kconfig +2 -0
index c2bac9c..34aafec 100644
@@ -853,6 +853,19 @@ config NETFILTER_XT_MATCH_U32
 
 	  Details and examples are in the kernel module source.
 
+config NETFILTER_XT_MATCH_OSF
+	tristate 'Passive OS fingerprint match support'

Just a minor thing, could you add "osf" somewhere in the short-text so that people instantly know the name of the module (without having to consult the help text), like the other entries. For example

tristate '"osf" Passive OS fingerprint match'

net/netfilter/Kconfig config NETFILTER_XT_MATCH_U32 · line 862

Quoted 5 lines
net/netfilter/Kconfig +5 -0
+	depends on NETFILTER_ADVANCED
+	help
+	  Passive OS fingerprint matching module. Allows to passively match
+	  remote operation system analyzing incoming TCP packets with SYN
+	  bit set.

A bit of rewording I suggest.

This option selects the Passive OS Fingerprinting match module that allows to passively match the
remote operating system by analyzing incoming TCP SYN packets.

net/netfilter/Kconfig config NETFILTER_XT_MATCH_U32 · line 865

Quoted 3 lines
net/netfilter/Kconfig +3 -0
+
+	  You should download and install rule loading software from
+	  http://www.ioremap.net/projects/osf

should -> can

Rules and loading software can be downloaded from
http://ioremap.net/projects/osf/

Quoted 9 lines
net/netfilter/Makefile +0 -0
index da3d909..ec3a6e0 100644
@@ -89,6 +89,7 @@ obj-$(CONFIG_NETFILTER_XT_MATCH_STRING) += xt_string.o

obj-$(CONFIG_NETFILTER_XT_MATCH_TCPMSS) += xt_tcpmss.o obj-$(CONFIG_NETFILTER_XT_MATCH_TIME) += xt_time.o obj-$(CONFIG_NETFILTER_XT_MATCH_U32) += xt_u32.o +obj-$(CONFIG_NETFILTER_XT_MATCH_OSF) += xt_osf.o

Please have the list sorted alphabetically. (This keeps merge conflicts down because everybody does not try to append at the end.)

net/netfilter/xt_osf.c line 86

Quoted 13 lines
net/netfilter/xt_osf.c new file +8 -0
new file mode 100644
index 0000000..f8b5a16
+
+static void ipt_osf_send_connector(struct ipt_osf_user_finger *f,
+				   const struct sk_buff *skb)
+{
+	struct ipt_osf_message *msg = &per_cpu(ipt_osf_mbuf, smp_processor_id());
+	struct ipt_osf_nlmsg *data = &msg->nlmsg;
+	struct iphdr *iph = ip_hdr(skb);
+	struct tcphdr *tcph = tcp_hdr(skb);

iph and tcph are only used in memcpy and could be directly substituted:

	memcpy(&data->ip, ip_hdr(skb), ...)
	memcpy(&data->tcp, tcp_hdr(skb), ...)

net/netfilter/xt_osf.c line 94

Quoted 7 lines
net/netfilter/xt_osf.c new file +7 -0
+	memcpy(&msg->cmsg.id, &cn_osf_id, sizeof(struct cn_msg));
+	msg->cmsg.seq = osf_seq++;
+	msg->cmsg.ack = 0;
+	msg->cmsg.len = sizeof(struct ipt_osf_nlmsg);
+
+	memcpy(&data->f, f, sizeof(struct ipt_osf_user_finger));
+	memcpy(&data->ip, iph, sizeof(struct iphdr));

xt_osf does not look at IP options, is that so?

net/netfilter/xt_osf.c line 95

Quoted 1 line
net/netfilter/xt_osf.c new file +1 -0
+	memcpy(&data->tcp, tcph, sizeof(struct tcphdr));

xt_osf does not look at TCP options, is not it?

Note that you cannot directly use tcp_hdr(skb) as the skb->transport_header has not yet been initialized (it still points at skb->network_header) because the packet has not yet been seen by the next handler. This affects PREROUTING, and INPUT chains (and BROUTING, for ebtables people, but irrelevant here).

The packet may also be fragmented or non-linear.

You actually do the right thing in ipt_osf_match_packet.

Quoted 1 line

+static inline int ipt_osf_ttl(const struct sk_buff *skb, struct ipt_osf_info *info,

const struct ipt_osf_info *

net/netfilter/xt_osf.c line 103

Quoted 3 lines
net/netfilter/xt_osf.c new file +3 -0
+			    unsigned char f_ttl)
+{
+	struct iphdr *ip = ip_hdr(skb);

const struct iphdr *iph

net/netfilter/xt_osf.c line 107

Quoted 3 lines
net/netfilter/xt_osf.c new file +3 -0
+	if (info->flags & IPT_OSF_TTL) {
+		if (info->ttl == IPT_OSF_TTL_TRUE)
+			return (ip->ttl == f_ttl);

redundant pair of ()

Quoted 12 lines
net/netfilter/xt_osf.c new file +12 -0
+		if (info->ttl == IPT_OSF_TTL_NOCHECK)
+			return 1;
+		else if (ip->ttl <= f_ttl)
+			return 1;
+		else {
+			struct in_device *in_dev = in_dev_get(skb->dev);
+			int ret = 0;
+
+			for_ifa(in_dev) {
+				if (inet_ifa_match(ip->saddr, ifa)) {
+					ret = (ip->ttl == f_ttl);
+					break;

net/netfilter/xt_osf.c line 129

Quoted 10 lines
net/netfilter/xt_osf.c new file +10 -0
+				}
+			}
+			endfor_ifa(in_dev);
+
+			in_dev_put(in_dev);
+			return ret;
+		}
+	}
+	
+	return (ip->ttl == f_ttl);

redundant pair of ()

net/netfilter/xt_osf.c line 135

Quoted 6 lines
net/netfilter/xt_osf.c new file +6 -0
+}
+
+static bool ipt_osf_match_packet(const struct sk_buff *skb,
+		const struct xt_match_param *p)
+{
+	struct ipt_osf_info *info = (struct ipt_osf_info *)p->matchinfo;

no cast needed. const advised.

net/netfilter/xt_osf.c line 137

Quoted 2 lines
net/netfilter/xt_osf.c new file +2 -0
+	struct iphdr *ip;
+	struct tcphdr _tcph, *tcp;

const

net/netfilter/xt_osf.c line 142

Quoted 5 lines
net/netfilter/xt_osf.c new file +5 -0
+	int fmatch = FMATCH_WRONG, fcount = 0;
+	unsigned int optsize = 0, check_WSS = 0;
+	u16 window, totlen, mss = 0;
+	bool df;
+	unsigned char *optp = NULL, *_optp = NULL;

const

net/netfilter/xt_osf.c line 146

Quoted 5 lines
net/netfilter/xt_osf.c new file +4 -0
+	unsigned char opts[MAX_IPOPTLEN];
 
+	struct ipt_osf_finger *kf;
+	struct ipt_osf_user_finger *f;
+	struct ipt_osf_finger_storage *st;

three const, as far as possible

net/netfilter/xt_osf.c line 153

Quoted 7 lines
net/netfilter/xt_osf.c new file +7 -0
+
+	if (!info)
+		return false;
+
+	ip = ip_hdr(skb);
+	if (!ip)
+		return false;

ip_hdr is always returning non-NULL anyway.

net/netfilter/xt_osf.c line 225

Quoted 5 lines
net/netfilter/xt_osf.c new file +5 -0
+
+			for (optnum = 0; optnum < f->opt_num; ++optnum) {
+				if (f->opt[optnum].kind == (*optp)) {
+					__u32 len = f->opt[optnum].length;
+					__u8 *optend = optp + len;

const

Quoted 5 lines

[...]

net/netfilter/xt_osf.c new file +4 -0
+	}
+
+	if (fcount)
+		fmatch = FMATCH_OK;

net/netfilter/xt_osf.c line 349

Quoted 10 lines
net/netfilter/xt_osf.c new file +10 -0
+
+	return fmatch == FMATCH_OK;
+}
+
+static struct xt_match ipt_osf_match = {
+	.name 		= "osf",
+	.revision	= 0,
+	.family		= NFPROTO_IPV4,
+	.proto		= IPPROTO_TCP,
+	.hooks      	= (1 << NF_INET_LOCAL_IN) | (1 << NF_INET_PRE_ROUTING),

Would allow FORWARD work?

net/netfilter/xt_osf.c line 354

Quoted 5 lines
net/netfilter/xt_osf.c new file +5 -0
+	.match 		= ipt_osf_match_packet,
+	.matchsize	= sizeof(struct ipt_osf_info),
+	.me		= THIS_MODULE,
+};
+

-Jan

#3 lore raw

Hi Jan.

Thanks for the review, I will incorporate the changes and respin it soon.

Signature · 1 line
	Evgeniy Polyakov
#4 lore raw

Re: Passive OS fingerprint xtables match (iptables)

In reply to Evgeniy Polyakov
Quoted 5 lines

static void osf_init(struct xt_entry_match *m)
{
}

You can remove this function, as it does no initialization. Implicit NULL in struct xtables_match is handled.

Quoted 6 lines
		case '2': /* --ttl */
			if (*flags & IPT_OSF_TTL)
				exit_error(PARAMETER_PROBLEM, "Can't specify multiple ttl parameter");
			*flags |= IPT_OSF_TTL;
			info->flags |= IPT_OSF_TTL;
			info->ttl = atoi(argv[optind-1]);

This would allow specifying --ttl 12345.

Use of xtables_strtoui (v1.4.3-rc1+git) for bounds checking, and use of optarg seem beneficial:

	unsigned int num;

	if (!xtables_strtoui(optarg, NULL, &num, 0, UINT8_MAX))
		exit_error(PARAMETER_PROBLEM, "*shrug*");
	info->ttl = num;
Quoted 5 lines
		case '3': /* --log */
			if (*flags & IPT_OSF_LOG)
				exit_error(PARAMETER_PROBLEM, "Can't specify multiple log parameter");
			*flags |= IPT_OSF_LOG;
			info->loglevel = atoi(argv[optind-1]);

similarly.

Quoted 12 lines
			info->flags |= IPT_OSF_LOG;
			break;
[...]
	return 1;
}

static void osf_final_check(unsigned int flags)

{
	if (!flags)
		exit_error(PARAMETER_PROBLEM, "OS fingerprint match: You must specify `--genre'");
}
Quoted 13 lines
static struct xtables_match osf_match = {
	.name		= "osf",
	.version	= XTABLES_VERSION,
	.size		= XT_ALIGN(sizeof(struct ipt_osf_info)),
	.userspacesize	= XT_ALIGN(sizeof(struct ipt_osf_info)),
	.help		= osf_help,
	.init		= osf_init,
	.parse		= osf_parse,
	.print		= osf_print,
	.final_check	= osf_final_check,
	.save		= osf_save,
	.extra_opts	= osf_opts
};

Needs .family = NFPROTO_IPV4.

#5 lore raw
In reply to Evgeniy Polyakov
Quoted 23 lines · Evgeniy Polyakov
net/netfilter/xt_osf.c new file +21 -0
+struct ipt_osf_finger {
+	struct rcu_head			rcu_head;
+	struct list_head		finger_entry;
+	struct ipt_osf_user_finger	finger;
+};
 
+struct ipt_osf_finger_storage
+{
+	struct list_head		finger_list;
+	spinlock_t			finger_lock;
+};
+
+/*
+ * Indexed by dont-fragment bit.
+ * It is the only constant value in the fingerprint.
+ */
+struct ipt_osf_finger_storage ipt_osf_fingers[2];
+
+static DEFINE_PER_CPU(struct ipt_osf_message, ipt_osf_mbuf);
 
+static void __exit ipt_osf_fini(void)
+{
+	struct ipt_osf_finger *f;

net/netfilter/xt_osf.c

Quoted 10 lines
net/netfilter/xt_osf.c new file +10 -0
+	int i;
+
+	cn_del_callback(&cn_osf_id);
+	xt_unregister_match(&ipt_osf_match);
+
+	rcu_read_lock();
+	for (i=0; i<ARRAY_SIZE(ipt_osf_fingers); ++i) {
+		struct ipt_osf_finger_storage *st = &ipt_osf_fingers[i];
+
+		list_for_each_entry_rcu(f, &st->finger_list, finger_entry) {

spin_lock(&st->finger_lock); //???

net/netfilter/xt_osf.c line 459

Quoted 1 line
net/netfilter/xt_osf.c new file +1 -0
+			list_del_rcu(&f->finger_entry);

spin_unlock(&st->finger_lock);

net/netfilter/xt_osf.c line 463

Quoted 4 lines
net/netfilter/xt_osf.c new file +4 -0
+			call_rcu(&f->rcu_head, ipt_osf_finger_free_rcu);
+		}
+	}
+	rcu_read_unlock();

Should the list_del_rcu() not be protected by a spinlock?

net/netfilter/xt_osf.c line 465

Quoted 1 line
net/netfilter/xt_osf.c new file +1 -0
+	rcu_barrier();

In some of my code I call synchronize_net(), is it enough to call rcu_barrier()?

What is the difference between:

synchronize_rcu()
synchronize_net()
rcu_barrier()

Hilsen
Jesper Brouer

--
-------------------------------------------------------------------
MSc. Master of Computer Science
Dept. of Computer Science, University of Copenhagen Author of http://www.adsl-optimizer.dk
-------------------------------------------------------------------

#6 lore raw

Hi.

Quoted 18 lines · Jesper Dangaard Brouer

net/netfilter/xt_osf.c line 458

Quoted 13 lines
net/netfilter/xt_osf.c new file +13 -0
+static void __exit ipt_osf_fini(void)
+{
+	struct ipt_osf_finger *f;
+	int i;
+
+	cn_del_callback(&cn_osf_id);
+	xt_unregister_match(&ipt_osf_match);
+
+	rcu_read_lock();
+	for (i=0; i<ARRAY_SIZE(ipt_osf_fingers); ++i) {
+		struct ipt_osf_finger_storage *st = &ipt_osf_fingers[i];
+
+		list_for_each_entry_rcu(f, &st->finger_list, finger_entry) {

spin_lock(&st->finger_lock); //???

net/netfilter/xt_osf.c line 459

Quoted 1 line
net/netfilter/xt_osf.c new file +1 -0
+			list_del_rcu(&f->finger_entry);

spin_unlock(&st->finger_lock);

Quoted 6 lines

net/netfilter/xt_osf.c line 463

Quoted 4 lines
net/netfilter/xt_osf.c new file +4 -0
+			call_rcu(&f->rcu_head, ipt_osf_finger_free_rcu);
+		}
+	}
+	rcu_read_unlock();

Should the list_del_rcu() not be protected by a spinlock?

Not required at this place - all users are already unregistered and no code can access this list except module exit path.

Quoted 4 lines

net/netfilter/xt_osf.c line 465

Quoted 1 line
net/netfilter/xt_osf.c new file +1 -0
+	rcu_barrier();

In some of my code I call synchronize_net(), is it enough to call rcu_barrier()?

It is enough here, rcu_barrier() will wait until all scheduled call_rcu() are completed, that's what we need. But in some cases we should only wait for the whole grace period to elapse, then one has to use synchronize_rcu() and friends. rcu_barrier() will wait for the callbacks to be executed, while they are executed after grace period has elapsed, so it implicitly includes synchronize_rcu(), but effectively they are the same: both functions register rcu callback and wait for the completion, rcu_barrier() is a bit more enhanced, since it has several types.

Quoted 4 lines

What is the difference between:

synchronize_rcu()
synchronize_net()

Those are essentially the same - synchronize_net() has additional might_sleep() call. Both will wait until grace period elapced - i.e. all currently RCU protected sections completed.

Quoted 1 line

rcu_barrier()

It will wait until all scheduled rcu callbacks are executed.

So from the description they look different, but implementation suggestes that effectively they are the same, except that there are a bit different invocation types for the barrier.

Signature · 1 line
	Evgeniy Polyakov
#7 lore raw
In reply to Jesper Dangaard Brouer
Quoted 12 lines · Jesper Dangaard Brouer

net/netfilter/xt_osf.c line 463

Quoted 4 lines · Evgeniy Polyakov
net/netfilter/xt_osf.c new file +4 -0
+            call_rcu(&f->rcu_head, ipt_osf_finger_free_rcu);
+        }
+    }
+    rcu_read_unlock();

Should the list_del_rcu() not be protected by a spinlock?

Quoted 1 line

+ rcu_barrier();

Quoted 7 lines

In some of my code I call synchronize_net(), is it enough to call rcu_barrier()?

What is the difference between:

synchronize_rcu()
synchronize_net()
rcu_barrier()

synchronize_net() is just a call to synchronize_rcu(), so their functionality is equivalent. synchronize_net() is however only supposed to synchronize with RX packet processing, which is usually not enough for netfilter. So I prefer synchronize_rcu() for clarity.

#8 lore raw
In reply to Evgeniy Polyakov
Quoted 11 lines · Evgeniy Polyakov

Hi.

Passive OS fingerprinting netfilter module allows to passively detect remote OS and perform various netfilter actions based on that knowledge. This module compares some data (WS, MSS, options and it's order, ttl, df and others) from packets with SYN bit set with dynamically loaded OS fingerprints.

Fingerprint matching rules can be downloaded from OpenBSD source tree and loaded via netlink connector into the kernel via special util found in archive. It will also listen for events about matching packets.

I like this feature. We have nfnetlink so I don't see why we should use the netlink connector instead.

BTW, is there any difference with regards to userspace p0f apart from having this integrated into iptables?

Signature · 1 line
"Los honestos son inadaptados sociales" -- Les Luthiers
#9 lore raw

Hi Pablo.

Quoted 6 lines · Pablo Neira Ayuso
Quoted 3 lines

Fingerprint matching rules can be downloaded from OpenBSD source tree and loaded via netlink connector into the kernel via special util found in archive. It will also listen for events about matching packets.

I like this feature. We have nfnetlink so I don't see why we should use the netlink connector instead.

OSF exists about 6 years already, netlink configuration was added in 2005, I do not remember if nfnetlink existed those days (IIRC it did not, since I reused ULOG netlink first), right now I just cleanup what was written before.

Quoted 2 lines

BTW, is there any difference with regards to userspace p0f apart from having this integrated into iptables?

There should be no major differences, there are some tweaks for the MTU comparison, maybe something else.

Signature · 1 line
	Evgeniy Polyakov
#10 lore raw
Quoted 8 lines · Evgeniy Polyakov

Hi Pablo.

Quoted 5 lines · Pablo Neira Ayuso
Quoted 3 lines

Fingerprint matching rules can be downloaded from OpenBSD source tree and loaded via netlink connector into the kernel via special util found in archive. It will also listen for events about matching packets.

I like this feature. We have nfnetlink so I don't see why we should use the netlink connector instead.

I fully agree.

Quoted 4 lines

OSF exists about 6 years already, netlink configuration was added in 2005, I do not remember if nfnetlink existed those days (IIRC it did not, since I reused ULOG netlink first), right now I just cleanup what was written before.

We do have nfnetlink today however, so this argument does no longer apply. I don't mind the order in which things are fixed up of course, but before merging, it needs to be converted to nfnetlink.