Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1467005 > unrolled thread

[PATCH 3.10 007/180] netfilter: x_tables: validate targets of jumps

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH 3.10 007/180] netfilter: x_tables: validate targets of jumps Willy Tarreau <w@1wt.eu> - 2016-08-21 17:40 +0200
    Re: [PATCH 3.10 007/180] netfilter: x_tables: validate targets of  jumps Willy Tarreau <w@1wt.eu> - 2016-08-21 22:00 +0200
      Re: [PATCH 3.10 007/180] netfilter: x_tables: validate targets of  jumps Florian Westphal <fw@strlen.de> - 2016-08-21 22:20 +0200
        Re: [PATCH 3.10 007/180] netfilter: x_tables: validate targets of  jumps Willy Tarreau <w@1wt.eu> - 2016-08-21 23:30 +0200

#1467005 — [PATCH 3.10 007/180] netfilter: x_tables: validate targets of jumps

From: Florian Westphal <fw@strlen.de>

commit 36472341017529e2b12573093cc0f68719300997 upstream.

When we see a jump also check that the offset gets us to beginning of
a rule (an ipt_entry).

The extra overhead is negible, even with absurd cases.

300k custom rules, 300k jumps to 'next' user chain:
[ plus one jump from INPUT to first userchain ]:

Before:
real    0m24.874s
user    0m7.532s
sys     0m16.076s

After:
real    0m27.464s
user    0m7.436s
sys     0m18.840s

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
---
 net/ipv4/netfilter/arp_tables.c | 16 ++++++++++++++++
 net/ipv4/netfilter/ip_tables.c  | 16 ++++++++++++++++
 net/ipv6/netfilter/ip6_tables.c | 16 ++++++++++++++++
 3 files changed, 48 insertions(+)

diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
index 7460b7b..b2adc4b 100644
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -358,6 +358,18 @@ static inline bool unconditional(const struct arpt_entry *e)
 	       memcmp(&e->arp, &uncond, sizeof(uncond)) == 0;
 }
 
+static bool find_jump_target(const struct xt_table_info *t,
+			     const struct arpt_entry *target)
+{
+	struct arpt_entry *iter;
+
+	xt_entry_foreach(iter, t->entries, t->size) {
+		 if (iter == target)
+			return true;
+	}
+	return false;
+}
+
 /* Figures out from what hook each rule can be called: returns 0 if
  * there are loops.  Puts hook bitmask in comefrom.
  */
@@ -451,6 +463,10 @@ static int mark_source_chains(const struct xt_table_info *newinfo,
 					/* This a jump; chase it. */
 					duprintf("Jump rule %u -> %u\n",
 						 pos, newpos);
+					e = (struct arpt_entry *)
+						(entry0 + newpos);
+					if (!find_jump_target(newinfo, e))
+						return 0;
 				} else {
 					/* ... this is a fallthru */
 					newpos = pos + e->next_offset;
diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c
index 8fc22ee..eea26fc 100644
--- a/net/ipv4/netfilter/ip_tables.c
+++ b/net/ipv4/netfilter/ip_tables.c
@@ -434,6 +434,18 @@ ipt_do_table(struct sk_buff *skb,
 #endif
 }
 
+static bool find_jump_target(const struct xt_table_info *t,
+			     const struct ipt_entry *target)
+{
+	struct ipt_entry *iter;
+
+	xt_entry_foreach(iter, t->entries, t->size) {
+		 if (iter == target)
+			return true;
+	}
+	return false;
+}
+
 /* Figures out from what hook each rule can be called: returns 0 if
    there are loops.  Puts hook bitmask in comefrom. */
 static int
@@ -531,6 +543,10 @@ mark_source_chains(const struct xt_table_info *newinfo,
 					/* This a jump; chase it. */
 					duprintf("Jump rule %u -> %u\n",
 						 pos, newpos);
+					e = (struct ipt_entry *)
+						(entry0 + newpos);
+					if (!find_jump_target(newinfo, e))
+						return 0;
 				} else {
 					/* ... this is a fallthru */
 					newpos = pos + e->next_offset;
diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c
index 63f7876..5eeec96 100644
--- a/net/ipv6/netfilter/ip6_tables.c
+++ b/net/ipv6/netfilter/ip6_tables.c
@@ -444,6 +444,18 @@ ip6t_do_table(struct sk_buff *skb,
 #endif
 }
 
+static bool find_jump_target(const struct xt_table_info *t,
+			     const struct ip6t_entry *target)
+{
+	struct ip6t_entry *iter;
+
+	xt_entry_foreach(iter, t->entries, t->size) {
+		 if (iter == target)
+			return true;
+	}
+	return false;
+}
+
 /* Figures out from what hook each rule can be called: returns 0 if
    there are loops.  Puts hook bitmask in comefrom. */
 static int
@@ -541,6 +553,10 @@ mark_source_chains(const struct xt_table_info *newinfo,
 					/* This a jump; chase it. */
 					duprintf("Jump rule %u -> %u\n",
 						 pos, newpos);
+					e = (struct ip6t_entry *)
+						(entry0 + newpos);
+					if (!find_jump_target(newinfo, e))
+						return 0;
 				} else {
 					/* ... this is a fallthru */
 					newpos = pos + e->next_offset;
-- 
2.8.0.rc2.1.gbe9624a

[toc] | [next] | [standalone]

#1467229 — Re: [PATCH 3.10 007/180] netfilter: x_tables: validate targets of jumps

On Sun, Aug 21, 2016 at 05:28:57PM +0200, Willy Tarreau wrote:
> From: Florian Westphal <fw@strlen.de>
> 
> commit 36472341017529e2b12573093cc0f68719300997 upstream.
> 
> When we see a jump also check that the offset gets us to beginning of
> a rule (an ipt_entry).
(...)

Sorry, this patch is bad and causes performance issues, I didn't notice
that 3.14 had a different version, below. Thanks to Jay for reporting
the problem to me. I'll push 3.10.103-rc2 after some cool down period.

Willy

----

From 3c079e0993ca3bdd94dddd028ee04c85911aa666 Mon Sep 17 00:00:00 2001
From: Florian Westphal <fw@strlen.de>
Date: Fri, 1 Apr 2016 14:17:22 +0200
Subject: netfilter: x_tables: validate targets of jumps

commit 36472341017529e2b12573093cc0f68719300997 upstream.

When we see a jump also check that the offset gets us to beginning of
a rule (an ipt_entry).

The extra overhead is negible, even with absurd cases.

300k custom rules, 300k jumps to 'next' user chain:
[ plus one jump from INPUT to first userchain ]:

Before:
real    0m24.874s
user    0m7.532s
sys     0m16.076s

After:
real    0m27.464s
user    0m7.436s
sys     0m18.840s

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
[wt: pick the patch from 3.14, it slightly differs to address the issue
 described here : https://patchwork.ozlabs.org/patch/650121/]
Signed-off-by: Willy Tarreau <w@1wt.eu>
---
 net/ipv4/netfilter/arp_tables.c | 17 +++++++++++++++++
 net/ipv4/netfilter/ip_tables.c  | 17 +++++++++++++++++
 net/ipv6/netfilter/ip6_tables.c | 17 +++++++++++++++++
 3 files changed, 51 insertions(+)

diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
index 7460b7b..473ec55 100644
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -358,6 +358,19 @@ static inline bool unconditional(const struct arpt_entry *e)
 	       memcmp(&e->arp, &uncond, sizeof(uncond)) == 0;
 }
 
+static bool find_jump_target(const struct xt_table_info *t,
+			     const void *entry0,
+			     const struct arpt_entry *target)
+{
+	struct arpt_entry *iter;
+
+	xt_entry_foreach(iter, entry0, t->size) {
+		 if (iter == target)
+			return true;
+	}
+	return false;
+}
+
 /* Figures out from what hook each rule can be called: returns 0 if
  * there are loops.  Puts hook bitmask in comefrom.
  */
@@ -451,6 +464,10 @@ static int mark_source_chains(const struct xt_table_info *newinfo,
 					/* This a jump; chase it. */
 					duprintf("Jump rule %u -> %u\n",
 						 pos, newpos);
+					e = (struct arpt_entry *)
+						(entry0 + newpos);
+					if (!find_jump_target(newinfo, entry0, e))
+						return 0;
 				} else {
 					/* ... this is a fallthru */
 					newpos = pos + e->next_offset;
diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c
index 8fc22ee..dba9d80 100644
--- a/net/ipv4/netfilter/ip_tables.c
+++ b/net/ipv4/netfilter/ip_tables.c
@@ -434,6 +434,19 @@ ipt_do_table(struct sk_buff *skb,
 #endif
 }
 
+static bool find_jump_target(const struct xt_table_info *t,
+			     const void *entry0,
+			     const struct ipt_entry *target)
+{
+	struct ipt_entry *iter;
+
+	xt_entry_foreach(iter, entry0, t->size) {
+		 if (iter == target)
+			return true;
+	}
+	return false;
+}
+
 /* Figures out from what hook each rule can be called: returns 0 if
    there are loops.  Puts hook bitmask in comefrom. */
 static int
@@ -531,6 +544,10 @@ mark_source_chains(const struct xt_table_info *newinfo,
 					/* This a jump; chase it. */
 					duprintf("Jump rule %u -> %u\n",
 						 pos, newpos);
+					e = (struct ipt_entry *)
+						(entry0 + newpos);
+					if (!find_jump_target(newinfo, entry0, e))
+						return 0;
 				} else {
 					/* ... this is a fallthru */
 					newpos = pos + e->next_offset;
diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c
index 63f7876..97a8d25 100644
--- a/net/ipv6/netfilter/ip6_tables.c
+++ b/net/ipv6/netfilter/ip6_tables.c
@@ -444,6 +444,19 @@ ip6t_do_table(struct sk_buff *skb,
 #endif
 }
 
+static bool find_jump_target(const struct xt_table_info *t,
+			     const void *entry0,
+			     const struct ip6t_entry *target)
+{
+	struct ip6t_entry *iter;
+
+	xt_entry_foreach(iter, entry0, t->size) {
+		 if (iter == target)
+			return true;
+	}
+	return false;
+}
+
 /* Figures out from what hook each rule can be called: returns 0 if
    there are loops.  Puts hook bitmask in comefrom. */
 static int
@@ -541,6 +554,10 @@ mark_source_chains(const struct xt_table_info *newinfo,
 					/* This a jump; chase it. */
 					duprintf("Jump rule %u -> %u\n",
 						 pos, newpos);
+					e = (struct ip6t_entry *)
+						(entry0 + newpos);
+					if (!find_jump_target(newinfo, entry0, e))
+						return 0;
 				} else {
 					/* ... this is a fallthru */
 					newpos = pos + e->next_offset;
-- 
2.8.0.rc2.1.gbe9624a

[toc] | [prev] | [next] | [standalone]

#1467237 — Re: [PATCH 3.10 007/180] netfilter: x_tables: validate targets of jumps

Willy Tarreau <w@1wt.eu> wrote:
> On Sun, Aug 21, 2016 at 05:28:57PM +0200, Willy Tarreau wrote:
> > From: Florian Westphal <fw@strlen.de>
> > 
> > commit 36472341017529e2b12573093cc0f68719300997 upstream.
> > 
> > When we see a jump also check that the offset gets us to beginning of
> > a rule (an ipt_entry).
> (...)
> 
> Sorry, this patch is bad and causes performance issues, I didn't notice
> that 3.14 had a different version, below. Thanks to Jay for reporting
> the problem to me. I'll push 3.10.103-rc2 after some cool down period.

Please either hold this one back or also queue
commit f4dc77713f8016d2e8a3295e1c9c53a21f296def
netfilter: x_tables: speed up jump target validation

It supersedes this one.

> +static bool find_jump_target(const struct xt_table_info *t,
> +			     const void *entry0,
> +			     const struct arpt_entry *target)
> +{
> +	struct arpt_entry *iter;
> +
> +	xt_entry_foreach(iter, entry0, t->size) {
> +		 if (iter == target)
> +			return true;
> +	}

This linear search works but its too slow with large rule sets.

[toc] | [prev] | [next] | [standalone]

#1467258 — Re: [PATCH 3.10 007/180] netfilter: x_tables: validate targets of jumps

On Sun, Aug 21, 2016 at 10:16:16PM +0200, Florian Westphal wrote:
> Willy Tarreau <w@1wt.eu> wrote:
> > On Sun, Aug 21, 2016 at 05:28:57PM +0200, Willy Tarreau wrote:
> > > From: Florian Westphal <fw@strlen.de>
> > > 
> > > commit 36472341017529e2b12573093cc0f68719300997 upstream.
> > > 
> > > When we see a jump also check that the offset gets us to beginning of
> > > a rule (an ipt_entry).
> > (...)
> > 
> > Sorry, this patch is bad and causes performance issues, I didn't notice
> > that 3.14 had a different version, below. Thanks to Jay for reporting
> > the problem to me. I'll push 3.10.103-rc2 after some cool down period.
> 
> Please either hold this one back or also queue
> commit f4dc77713f8016d2e8a3295e1c9c53a21f296def
> netfilter: x_tables: speed up jump target validation
> 
> It supersedes this one.

OK, I'm postponing it then, as the patch above is not yet in more recent
stable versions.

Thanks Florian!
Willy

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web