Skip to contentSkip to conversation

netfilter: fix ordering of jumpstack allocation and table update

PATCH Applied Will Deacon 3 participants 7 messages Oct 17, 2013 – Oct 22, 2013

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 netfilter: fix ordering of jumpstack allocation and table update

During kernel stability testing on an SMP ARMv7 system, Yalin Wang reported the following panic from the netfilter code:

1fe0: 0000001c 5e2d3b10 4007e779 4009e110 60000010 00000032 ff565656 ff545454 [<c06c48dc>] (ipt_do_table+0x448/0x584) from [<c0655ef0>] (nf_iterate+0x48/0x7c) [<c0655ef0>] (nf_iterate+0x48/0x7c) from [<c0655f7c>] (nf_hook_slow+0x58/0x104) [<c0655f7c>] (nf_hook_slow+0x58/0x104) from [<c0683bbc>] (ip_local_deliver+0x88/0xa8) [<c0683bbc>] (ip_local_deliver+0x88/0xa8) from [<c0683718>] (ip_rcv_finish+0x418/0x43c) [<c0683718>] (ip_rcv_finish+0x418/0x43c) from [<c062b1c4>] (__netif_receive_skb+0x4cc/0x598) [<c062b1c4>] (__netif_receive_skb+0x4cc/0x598) from [<c062b314>] (process_backlog+0x84/0x158) [<c062b314>] (process_backlog+0x84/0x158) from [<c062de84>] (net_rx_action+0x70/0x1dc) [<c062de84>] (net_rx_action+0x70/0x1dc) from [<c0088230>] (__do_softirq+0x11c/0x27c) [<c0088230>] (__do_softirq+0x11c/0x27c) from [<c008857c>] (do_softirq+0x44/0x50) [<c008857c>] (do_softirq+0x44/0x50) from [<c0088614>] (local_bh_enable_ip+0x8c/0xd0) [<c0088614>] (local_bh_enable_ip+0x8c/0xd0) from [<c06b0330>] (inet_stream_connect+0x164/0x298) [<c06b0330>] (inet_stream_connect+0x164/0x298) from [<c061d68c>] (sys_connect+0x88/0xc8) [<c061d68c>] (sys_connect+0x88/0xc8) from [<c000e340>] (ret_fast_syscall+0x0/0x30)

  Code: 2a000021 e59d2028 e59de01c e59f011c (e7824103)
  ---[ end trace da227214a82491bd ]---
  Kernel panic - not syncing: Fatal exception in interrupt

This comes about because CPU1 is executing xt_replace_table in response to a setsockopt syscall, resulting in:

	ret = xt_jumpstack_alloc(newinfo);
		--> newinfo->jumpstack = kzalloc(size, GFP_KERNEL);

	[...]

	table->private = newinfo;
	newinfo->initial_entries = private->initial_entries;

Meanwhile, CPU0 is handling the network receive path and ends up in ipt_do_table, resulting in:

	private = table->private;

	[...]

	jumpstack  = (struct ipt_entry **)private->jumpstack[cpu];

On weakly ordered memory architectures, the writes to table->private and newinfo->jumpstack from CPU1 can be observed out of order by CPU0. Furthermore, on architectures which don't respect ordering of address dependencies (i.e. Alpha), the reads from CPU0 can also be re-ordered.

This patch adds an smp_wmb() before the assignment to table->private (which is essentially publishing newinfo) to ensure that all writes to newinfo will be observed before plugging it into the table structure. A dependent-read barrier is also added on the consumer side, to ensure the same ordering requirement are also respected there.

---
 net/ipv4/netfilter/ip_tables.c | 5 +++++
 net/netfilter/x_tables.c       | 7 ++++++-
 2 files changed, 11 insertions(+), 1 deletion(-)
net/ipv4/netfilter/ip_tables.c +5 -0
index d23118d..3686458 100644
@@ -327,6 +327,11 @@ ipt_do_table(struct sk_buff *skb,
 	addend = xt_write_recseq_begin();
 	private = table->private;
 	cpu        = smp_processor_id();
+	/*
+	 * Ensure we load private members after we've fetched the base
+	 * pointer.
+	 */
+	smp_read_barrier_depends();
 	table_base = private->entries[cpu];
 	jumpstack  = (struct ipt_entry **)private->jumpstack[cpu];
 	stackptr   = per_cpu_ptr(private->stackptr, cpu);
net/netfilter/x_tables.c +6 -1
index 8b03028..227aa11 100644
@@ -845,8 +845,13 @@ xt_replace_table(struct xt_table *table,
 		return NULL;
 	}
 
-	table->private = newinfo;
 	newinfo->initial_entries = private->initial_entries;
+	/*
+	 * Ensure contents of newinfo are visible before assigning to
+	 * private.
+	 */
+	smp_wmb();
+	table->private = newinfo;
 
 	/*
 	 * Even though table entries have now been swapped, other CPU's
Signature · 1 line
1.8.2.2
Acked-by on v1 1/1
#1 lore raw
Quoted 15 lines · Will Deacon

During kernel stability testing on an SMP ARMv7 system, Yalin Wang reported the following panic from the netfilter code:

On weakly ordered memory architectures, the writes to table->private and newinfo->jumpstack from CPU1 can be observed out of order by CPU0. Furthermore, on architectures which don't respect ordering of address dependencies (i.e. Alpha), the reads from CPU0 can also be re-ordered.

This patch adds an smp_wmb() before the assignment to table->private (which is essentially publishing newinfo) to ensure that all writes to newinfo will be observed before plugging it into the table structure. A dependent-read barrier is also added on the consumer side, to ensure the same ordering requirement are also respected there.

Quoted 6 lines

---

Nice catch !

#2 lore raw
In reply to Will Deacon
Quoted 46 lines · Will Deacon

During kernel stability testing on an SMP ARMv7 system, Yalin Wang reported the following panic from the netfilter code:

1fe0: 0000001c 5e2d3b10 4007e779 4009e110 60000010 00000032 ff565656 ff545454 [<c06c48dc>] (ipt_do_table+0x448/0x584) from [<c0655ef0>] (nf_iterate+0x48/0x7c) [<c0655ef0>] (nf_iterate+0x48/0x7c) from [<c0655f7c>] (nf_hook_slow+0x58/0x104) [<c0655f7c>] (nf_hook_slow+0x58/0x104) from [<c0683bbc>] (ip_local_deliver+0x88/0xa8) [<c0683bbc>] (ip_local_deliver+0x88/0xa8) from [<c0683718>] (ip_rcv_finish+0x418/0x43c) [<c0683718>] (ip_rcv_finish+0x418/0x43c) from [<c062b1c4>] (__netif_receive_skb+0x4cc/0x598) [<c062b1c4>] (__netif_receive_skb+0x4cc/0x598) from [<c062b314>] (process_backlog+0x84/0x158) [<c062b314>] (process_backlog+0x84/0x158) from [<c062de84>] (net_rx_action+0x70/0x1dc) [<c062de84>] (net_rx_action+0x70/0x1dc) from [<c0088230>] (__do_softirq+0x11c/0x27c) [<c0088230>] (__do_softirq+0x11c/0x27c) from [<c008857c>] (do_softirq+0x44/0x50) [<c008857c>] (do_softirq+0x44/0x50) from [<c0088614>] (local_bh_enable_ip+0x8c/0xd0) [<c0088614>] (local_bh_enable_ip+0x8c/0xd0) from [<c06b0330>] (inet_stream_connect+0x164/0x298) [<c06b0330>] (inet_stream_connect+0x164/0x298) from [<c061d68c>] (sys_connect+0x88/0xc8) [<c061d68c>] (sys_connect+0x88/0xc8) from [<c000e340>] (ret_fast_syscall+0x0/0x30)

  Code: 2a000021 e59d2028 e59de01c e59f011c (e7824103)
  ---[ end trace da227214a82491bd ]---
  Kernel panic - not syncing: Fatal exception in interrupt

This comes about because CPU1 is executing xt_replace_table in response to a setsockopt syscall, resulting in:

	ret = xt_jumpstack_alloc(newinfo);
		--> newinfo->jumpstack = kzalloc(size, GFP_KERNEL);

	[...]

	table->private = newinfo;
	newinfo->initial_entries = private->initial_entries;

Meanwhile, CPU0 is handling the network receive path and ends up in ipt_do_table, resulting in:

	private = table->private;

	[...]

	jumpstack  = (struct ipt_entry **)private->jumpstack[cpu];

On weakly ordered memory architectures, the writes to table->private and newinfo->jumpstack from CPU1 can be observed out of order by CPU0. Furthermore, on architectures which don't respect ordering of address dependencies (i.e. Alpha), the reads from CPU0 can also be re-ordered.

Quoted 5 lines

This patch adds an smp_wmb() before the assignment to table->private (which is essentially publishing newinfo) to ensure that all writes to newinfo will be observed before plugging it into the table structure. A dependent-read barrier is also added on the consumer side, to ensure the same ordering requirement are also respected there.

We also need fixes for net/ipv6/netfilter/ip6_tables.c and net/ipv4/netfilter/arp_tables.c as well. Could you extend this patch and resend?

Thanks!

#3 lore raw

Hi Pablo,

Quoted 3 lines · Pablo Neira Ayuso
Quoted 2 lines · Will Deacon

During kernel stability testing on an SMP ARMv7 system, Yalin Wang reported the following panic from the netfilter code:

[...]

Quoted 9 lines
Quoted 5 lines

This patch adds an smp_wmb() before the assignment to table->private (which is essentially publishing newinfo) to ensure that all writes to newinfo will be observed before plugging it into the table structure. A dependent-read barrier is also added on the consumer side, to ensure the same ordering requirement are also respected there.

We also need fixes for net/ipv6/netfilter/ip6_tables.c and net/ipv4/netfilter/arp_tables.c as well. Could you extend this patch and resend?

Sure, I can try, but that's going to require a bit of time to sit down and look at the shared data, access order, dependencies etc. I'm currently preparing for Edinburgh, so it might be a while before I get a chance to extend this.

Will

#4 lore raw
Quoted 11 lines · Will Deacon

Hi Pablo,

Quoted 3 lines

We also need fixes for net/ipv6/netfilter/ip6_tables.c and net/ipv4/netfilter/arp_tables.c as well. Could you extend this patch and resend?

Sure, I can try, but that's going to require a bit of time to sit down and look at the shared data, access order, dependencies etc. I'm currently preparing for Edinburgh, so it might be a while before I get a chance to extend this.

That's basically same code copy/pasted, so it should be relatively easy.

author Tested-by Wang, Yalin
#5 lore raw
Quoted 8 lines · Eric Dumazet
Quoted 7 lines · Will Deacon

Hi Pablo,

Quoted 3 lines

We also need fixes for net/ipv6/netfilter/ip6_tables.c and net/ipv4/netfilter/arp_tables.c as well. Could you extend this patch and resend?

Quoted 6 lines
Quoted 4 lines

Sure, I can try, but that's going to require a bit of time to sit down and look at the shared data, access order, dependencies etc. I'm currently preparing for Edinburgh, so it might be a while before I get a chance to extend this.

That's basically same code copy/pasted, so it should be relatively easy.

Ok, I took a look and I think I see what you mean: there are just some additional consumers of the tables, so there aren't any additional writers afaict (at least, net/bridge/netfilter/ebtables.c uses rw locks so we don't have an issue there).

Attempt at an updated patch below.

Cheers,

Will

--->8

Quoted 1 line

From 128bff68debba70c3cec4b4ac01050d7e864aa83 Mon Sep 17 00:00:00 2001

From: Will Deacon <[email protected]>
Date: Thu, 17 Oct 2013 13:34:13 +0100
Subject: [PATCH v2] netfilter: fix ordering of jumpstack allocation and table
update

During kernel stability testing on an SMP ARMv7 system, Yalin Wang reported the following panic from the netfilter code:

1fe0: 0000001c 5e2d3b10 4007e779 4009e110 60000010 00000032 ff565656 ff545454 [<c06c48dc>] (ipt_do_table+0x448/0x584) from [<c0655ef0>] (nf_iterate+0x48/0x7c) [<c0655ef0>] (nf_iterate+0x48/0x7c) from [<c0655f7c>] (nf_hook_slow+0x58/0x104) [<c0655f7c>] (nf_hook_slow+0x58/0x104) from [<c0683bbc>] (ip_local_deliver+0x88/0xa8) [<c0683bbc>] (ip_local_deliver+0x88/0xa8) from [<c0683718>] (ip_rcv_finish+0x418/0x43c) [<c0683718>] (ip_rcv_finish+0x418/0x43c) from [<c062b1c4>] (__netif_receive_skb+0x4cc/0x598) [<c062b1c4>] (__netif_receive_skb+0x4cc/0x598) from [<c062b314>] (process_backlog+0x84/0x158) [<c062b314>] (process_backlog+0x84/0x158) from [<c062de84>] (net_rx_action+0x70/0x1dc) [<c062de84>] (net_rx_action+0x70/0x1dc) from [<c0088230>] (__do_softirq+0x11c/0x27c) [<c0088230>] (__do_softirq+0x11c/0x27c) from [<c008857c>] (do_softirq+0x44/0x50) [<c008857c>] (do_softirq+0x44/0x50) from [<c0088614>] (local_bh_enable_ip+0x8c/0xd0) [<c0088614>] (local_bh_enable_ip+0x8c/0xd0) from [<c06b0330>] (inet_stream_connect+0x164/0x298) [<c06b0330>] (inet_stream_connect+0x164/0x298) from [<c061d68c>] (sys_connect+0x88/0xc8) [<c061d68c>] (sys_connect+0x88/0xc8) from [<c000e340>] (ret_fast_syscall+0x0/0x30)

  Code: 2a000021 e59d2028 e59de01c e59f011c (e7824103)
  ---[ end trace da227214a82491bd ]---
  Kernel panic - not syncing: Fatal exception in interrupt

This comes about because CPU1 is executing xt_replace_table in response to a setsockopt syscall, resulting in:

	ret = xt_jumpstack_alloc(newinfo);
		--> newinfo->jumpstack = kzalloc(size, GFP_KERNEL);

	[...]

	table->private = newinfo;
	newinfo->initial_entries = private->initial_entries;

Meanwhile, CPU0 is handling the network receive path and ends up in ipt_do_table, resulting in:

	private = table->private;

	[...]

	jumpstack  = (struct ipt_entry **)private->jumpstack[cpu];

On weakly ordered memory architectures, the writes to table->private and newinfo->jumpstack from CPU1 can be observed out of order by CPU0. Furthermore, on architectures which don't respect ordering of address dependencies (i.e. Alpha), the reads from CPU0 can also be re-ordered.

This patch adds an smp_wmb() before the assignment to table->private (which is essentially publishing newinfo) to ensure that all writes to newinfo will be observed before plugging it into the table structure. A dependent-read barrier is also added on the consumer sides, to ensure the same ordering requirements are also respected there.

---
 net/ipv4/netfilter/arp_tables.c | 5 +++++
 net/ipv4/netfilter/ip_tables.c  | 5 +++++
 net/ipv6/netfilter/ip6_tables.c | 5 +++++
 net/netfilter/x_tables.c        | 7 ++++++-
 4 files changed, 21 insertions(+), 1 deletion(-)
net/ipv4/netfilter/arp_tables.c +5 -0
index 85a4f21..59da7cd 100644
@@ -271,6 +271,11 @@ unsigned int arpt_do_table(struct sk_buff *skb,
 	local_bh_disable();
 	addend = xt_write_recseq_begin();
 	private = table->private;
+	/*
+	 * Ensure we load private-> members after we've fetched the base
+	 * pointer.
+	 */
+	smp_read_barrier_depends();
 	table_base = private->entries[smp_processor_id()];
 
 	e = get_entry(table_base, private->hook_entry[hook]);
net/ipv4/netfilter/ip_tables.c +5 -0
index d23118d..718dfbd 100644
@@ -327,6 +327,11 @@ ipt_do_table(struct sk_buff *skb,
 	addend = xt_write_recseq_begin();
 	private = table->private;
 	cpu        = smp_processor_id();
+	/*
+	 * Ensure we load private-> members after we've fetched the base
+	 * pointer.
+	 */
+	smp_read_barrier_depends();
 	table_base = private->entries[cpu];
 	jumpstack  = (struct ipt_entry **)private->jumpstack[cpu];
 	stackptr   = per_cpu_ptr(private->stackptr, cpu);
net/ipv6/netfilter/ip6_tables.c +5 -0
index 44400c2..710238f 100644
@@ -349,6 +349,11 @@ ip6t_do_table(struct sk_buff *skb,
 	local_bh_disable();
 	addend = xt_write_recseq_begin();
 	private = table->private;
+	/*
+	 * Ensure we load private-> members after we've fetched the base
+	 * pointer.
+	 */
+	smp_read_barrier_depends();
 	cpu        = smp_processor_id();
 	table_base = private->entries[cpu];
 	jumpstack  = (struct ip6t_entry **)private->jumpstack[cpu];
net/netfilter/x_tables.c +6 -1
index 8b03028..227aa11 100644
@@ -845,8 +845,13 @@ xt_replace_table(struct xt_table *table,
 		return NULL;
 	}
 
-	table->private = newinfo;
 	newinfo->initial_entries = private->initial_entries;
+	/*
+	 * Ensure contents of newinfo are visible before assigning to
+	 * private.
+	 */
+	smp_wmb();
+	table->private = newinfo;
 
 	/*
 	 * Even though table entries have now been swapped, other CPU's
Signature · 1 line
1.8.3.2
maintainer Applied
#6 lore raw
Quoted 14 lines · Will Deacon
Quoted 13 lines · Eric Dumazet
Quoted 11 lines · Will Deacon

Hi Pablo,

Quoted 3 lines

We also need fixes for net/ipv6/netfilter/ip6_tables.c and net/ipv4/netfilter/arp_tables.c as well. Could you extend this patch and resend?

Sure, I can try, but that's going to require a bit of time to sit down and look at the shared data, access order, dependencies etc. I'm currently preparing for Edinburgh, so it might be a while before I get a chance to extend this.

Quoted 8 lines
Quoted 1 line

That's basically same code copy/pasted, so it should be relatively easy.

Ok, I took a look and I think I see what you mean: there are just some additional consumers of the tables, so there aren't any additional writers afaict (at least, net/bridge/netfilter/ebtables.c uses rw locks so we don't have an issue there).

Attempt at an updated patch below.

Applied, thanks Will.