Oops with latest (netfilter) nf-next tree, when unloading iptable_nat
Subsystems (3)
Hi Pablo,
I'm hitting this general protection fault, when unloading iptables_nat.
Git tree: git://1984.lsi.us.es/nf-next
At commit 0edd94887d19 (ipvs: use list_del_init instead of list_del/INIT_LIST_HEAD)
Notice, I'm not seeing this with net-next (at commit 9f00d9776bc5b)
[ 524.590171] general protection fault: 0000 [#1] SMP [ 524.591067] Modules linked in: netconsole ip_vs_lblc ip_vs_lc ip_vs_rr ip_vs libcrc32c ipt_MASQUERADE nf_nat_ipv4(-) nf_nat iptable_mangle xt_mark ip6table_mangle xt_LOG ip6table_filter ip6_tables virtio_balloon virtio_net [last unloaded: iptable_nat] [ 524.591067] CPU 0 [ 524.591067] Pid: 5842, comm: modprobe Not tainted 3.6.0-rc3-pablo-nf-next+ #1 Red Hat KVM [ 524.591067] RIP: 0010:[<ffffffffa002c2fd>] [<ffffffffa002c2fd>] nf_nat_proto_clean+0x6d/0xc0 [nf_nat] [ 524.591067] RSP: 0018:ffff880073203e18 EFLAGS: 00010246 [ 524.591067] RAX: 0000000000000000 RBX: ffff880077dff2c8 RCX: ffff8800797fab70 [ 524.591067] RDX: dead000000200200 RSI: ffff880073203e88 RDI: ffffffffa002f208 [ 524.591067] RBP: ffff880073203e28 R08: ffff880073202000 R09: 0000000000000000 [ 524.591067] R10: dead000000200200 R11: dead000000100100 R12: ffffffff81c6dc00 list corruption? ^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^ [ 524.591067] R13: ffff880078671640 R14: ffff880078671638 R15: ffff880073203e88 [ 524.591067] FS: 00007f04dc38b700(0000) GS:ffff88007cc00000(0000) knlGS:0000000000000000 [ 524.591067] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 524.591067] CR2: 00007f04dc398000 CR3: 0000000072238000 CR4: 00000000000006f0 [ 524.591067] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 524.591067] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 524.591067] Process modprobe (pid: 5842, threadinfo ffff880073202000, task ffff8800797fab70) [ 524.591067] Stack: [ 524.591067] ffff880073203e68 ffffffffa002c290 ffff880073203e78 ffffffff815614e3 [ 524.591067] ffffffff00000000 0000258d00000246 ffff880073203e68 ffffffff81c6dc00 [ 524.591067] ffff880073203e88 ffffffffa00358a0 0000000000000000 000000000040f5b0 [ 524.591067] Call Trace: [ 524.591067] [<ffffffffa002c290>] ? nf_nat_net_exit+0x50/0x50 [nf_nat] [ 524.591067] [<ffffffff815614e3>] nf_ct_iterate_cleanup+0xc3/0x170 [ 524.591067] [<ffffffffa002c54a>] nf_nat_l3proto_unregister+0x8a/0x100 [nf_nat] [ 524.591067] [<ffffffff812a0303>] ? compat_prepare_timeout+0x13/0xb0 [ 524.591067] [<ffffffffa0035848>] nf_nat_l3proto_ipv4_exit+0x10/0x23 [nf_nat_ipv4] [ 524.591067] [<ffffffff8109f4a5>] sys_delete_module+0x235/0x2b0 [ 524.591067] [<ffffffff810b8193>] ? __audit_syscall_entry+0x1b3/0x1f0 [ 524.591067] [<ffffffff810b8776>] ? __audit_syscall_exit+0x3e6/0x410 [ 524.591067] [<ffffffff816679e2>] system_call_fastpath+0x16/0x1b [ 524.591067] Code: 75 6c 0f b6 46 01 84 c0 74 05 3a 42 3e 75 5f 80 7e 02 00 74 41 48 c7 c7 08 f2 02 a0 e8 3d 3b 63 e1 48 8b 03 48 8b 53 08 48 85 c0 <48> 89 02 74 04 48 89 50 08 48 be 00 02 20 00 00 00 ad de 48 c7 [ 524.591067] RIP [<ffffffffa002c2fd>] nf_nat_proto_clean+0x6d/0xc0 [nf_nat] [ 524.591067] RSP <ffff880073203e18>
Signature · 5 lines
Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer
Florian Westphal
maintainerJesper Dangaard Brouer <[email protected]> wrote:
[ CC'd Patrick ]
Quoted 9 lines
I'm hitting this general protection fault, when unloading
iptables_nat.[ 524.591067] Pid: 5842, comm: modprobe Not tainted 3.6.0-rc3-pablo-nf-next+ #1 Red Hat KVM [ 524.591067] RIP: 0010:[<ffffffffa002c2fd>] [<ffffffffa002c2fd>] nf_nat_proto_clean+0x6d/0xc0 [nf_nat] [ 524.591067] RSP: 0018:ffff880073203e18 EFLAGS: 00010246 [ 524.591067] RAX: 0000000000000000 RBX: ffff880077dff2c8 RCX: ffff8800797fab70 [ 524.591067] RDX: dead000000200200 RSI: ffff880073203e88 RDI: ffffffffa002f208 [ 524.591067] RBP: ffff880073203e28 R08: ffff880073202000 R09: 0000000000000000 [ 524.591067] R10: dead000000200200 R11: dead000000100100 R12: ffffffff81c6dc00 list corruption? ^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^
Yep, looks like it.
Quoted 5 lines
[ 524.591067] [<ffffffffa002c290>] ? nf_nat_net_exit+0x50/0x50 [nf_nat] [ 524.591067] [<ffffffff815614e3>] nf_ct_iterate_cleanup+0xc3/0x170 [ 524.591067] [<ffffffffa002c54a>] nf_nat_l3proto_unregister+0x8a/0x100 [nf_nat] [ 524.591067] [<ffffffff812a0303>] ? compat_prepare_timeout+0x13/0xb0 [ 524.591067] [<ffffffffa0035848>] nf_nat_l3proto_ipv4_exit+0x10/0x23 [nf_nat_ipv4]
On module removal nf_nat_ipv4 calls nf_iterate_cleanup which invokes nf_nat_proto_clean() for each conntrack. That will then call hlist_del_rcu(&nat->bysource) using eachs conntracks nat ext area.
Problem is that nf_nat_proto_clean() is called multiple times for the same conntrack:
a) nf_ct_iterate_cleanup() returns each ct twice (origin, reply)
b) we call it both for l3 and for l4 protocol ids
We barf in hlist_del_rcu the 2nd time because ->pprev is poisoned.
This was introduced with the ipv6 nat patches.
net/netfilter/nf_nat_core.c +1 -1
@@ -487,7 +487,7 @@ static int nf_nat_proto_clean(struct nf_conn *i, void *data)if (clean->hash) {spin_lock_bh (&nf_nat_lock); -hlist_del_rcu (&nat->bysource); +hlist_del_init_rcu (&nat->bysource);spin_unlock_bh (&nf_nat_lock); }else {
Would probably avoid it. I guess it would be nicer to only call this once for each ct.
Patrick, any other idea?
Quoted 49 lines · Florian Westphal
Jesper Dangaard Brouer <[email protected]> wrote:
[ CC'd Patrick ]
Quoted 9 lines
I'm hitting this general protection fault, when unloading
iptables_nat.[ 524.591067] Pid: 5842, comm: modprobe Not tainted 3.6.0-rc3-pablo-nf-next+ #1 Red Hat KVM [ 524.591067] RIP: 0010:[<ffffffffa002c2fd>] [<ffffffffa002c2fd>] nf_nat_proto_clean+0x6d/0xc0 [nf_nat] [ 524.591067] RSP: 0018:ffff880073203e18 EFLAGS: 00010246 [ 524.591067] RAX: 0000000000000000 RBX: ffff880077dff2c8 RCX: ffff8800797fab70 [ 524.591067] RDX: dead000000200200 RSI: ffff880073203e88 RDI: ffffffffa002f208 [ 524.591067] RBP: ffff880073203e28 R08: ffff880073202000 R09: 0000000000000000 [ 524.591067] R10: dead000000200200 R11: dead000000100100 R12: ffffffff81c6dc00 list corruption? ^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^Yep, looks like it.
Quoted 5 lines
[ 524.591067] [<ffffffffa002c290>] ? nf_nat_net_exit+0x50/0x50 [nf_nat] [ 524.591067] [<ffffffff815614e3>] nf_ct_iterate_cleanup+0xc3/0x170 [ 524.591067] [<ffffffffa002c54a>] nf_nat_l3proto_unregister+0x8a/0x100 [nf_nat] [ 524.591067] [<ffffffff812a0303>] ? compat_prepare_timeout+0x13/0xb0 [ 524.591067] [<ffffffffa0035848>] nf_nat_l3proto_ipv4_exit+0x10/0x23 [nf_nat_ipv4]On module removal
nf_nat_ipv4callsnf_iterate_cleanupwhich invokesnf_nat_proto_clean()for each conntrack. That will then callhlist_del_rcu(&nat->bysource)using eachs conntracks nat ext area.Problem is that
nf_nat_proto_clean()is called multiple times for the same conntrack:
a)nf_ct_iterate_cleanup()returns each ct twice (origin, reply)
b) we call it both for l3 and for l4 protocol idsWe barf in
hlist_del_rcuthe 2nd time because ->pprev is poisoned.This was introduced with the ipv6 nat patches.
net/netfilter/nf_nat_core.c +1
-1@@ -487,7 +487,7 @@ static int nf_nat_proto_clean(struct nf_conn *i, void *data)if (clean->hash) {spin_lock_bh (&nf_nat_lock); -hlist_del_rcu (&nat->bysource); +hlist_del_init_rcu (&nat->bysource);spin_unlock_bh (&nf_nat_lock); }else {Would probably avoid it. I guess it would be nicer to only call this once for each ct.
Quoted 1 line
Patrick, any other idea?
I already discussed this with Florian (I've been having problems with two out of three of my email accounts this week... so I couldn't reply to this email in the mailing list).
We can add nf_nat_iterate_cleanup that can iterate over the NAT hashtable to replace current usage of nf_ct_iterate_cleanup.
Quoted 50 lines · Pablo Neira Ayuso
Quoted 49 lines · Florian Westphal
Jesper Dangaard Brouer <[email protected]> wrote:
[ CC'd Patrick ]
Quoted 9 lines
I'm hitting this general protection fault, when unloading
iptables_nat.[ 524.591067] Pid: 5842, comm: modprobe Not tainted 3.6.0-rc3-pablo-nf-next+ #1 Red Hat KVM [ 524.591067] RIP: 0010:[<ffffffffa002c2fd>] [<ffffffffa002c2fd>] nf_nat_proto_clean+0x6d/0xc0 [nf_nat] [ 524.591067] RSP: 0018:ffff880073203e18 EFLAGS: 00010246 [ 524.591067] RAX: 0000000000000000 RBX: ffff880077dff2c8 RCX: ffff8800797fab70 [ 524.591067] RDX: dead000000200200 RSI: ffff880073203e88 RDI: ffffffffa002f208 [ 524.591067] RBP: ffff880073203e28 R08: ffff880073202000 R09: 0000000000000000 [ 524.591067] R10: dead000000200200 R11: dead000000100100 R12: ffffffff81c6dc00 list corruption? ^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^Yep, looks like it.
Quoted 5 lines
[ 524.591067] [<ffffffffa002c290>] ? nf_nat_net_exit+0x50/0x50 [nf_nat] [ 524.591067] [<ffffffff815614e3>] nf_ct_iterate_cleanup+0xc3/0x170 [ 524.591067] [<ffffffffa002c54a>] nf_nat_l3proto_unregister+0x8a/0x100 [nf_nat] [ 524.591067] [<ffffffff812a0303>] ? compat_prepare_timeout+0x13/0xb0 [ 524.591067] [<ffffffffa0035848>] nf_nat_l3proto_ipv4_exit+0x10/0x23 [nf_nat_ipv4]On module removal
nf_nat_ipv4callsnf_iterate_cleanupwhich invokesnf_nat_proto_clean()for each conntrack. That will then callhlist_del_rcu(&nat->bysource)using eachs conntracks nat ext area.Problem is that
nf_nat_proto_clean()is called multiple times for the same conntrack:
a)nf_ct_iterate_cleanup()returns each ct twice (origin, reply)
b) we call it both for l3 and for l4 protocol idsWe barf in
hlist_del_rcuthe 2nd time because ->pprev is poisoned.This was introduced with the ipv6 nat patches.
net/netfilter/nf_nat_core.c +1
-1@@ -487,7 +487,7 @@ static int nf_nat_proto_clean(struct nf_conn *i, void *data)if (clean->hash) {spin_lock_bh (&nf_nat_lock); -hlist_del_rcu (&nat->bysource); +hlist_del_init_rcu (&nat->bysource);spin_unlock_bh (&nf_nat_lock); }else {Would probably avoid it. I guess it would be nicer to only call this once for each ct.
Quoted 8 lines
Quoted 1 line
Patrick, any other idea?
I already discussed this with Florian (I've been having problems with two out of three of my email accounts this week... so I couldn't reply to this email in the mailing list).
We can add
nf_nat_iterate_cleanupthat can iterate over the NAT hashtable to replace current usage ofnf_ct_iterate_cleanup.
Lets just bail out when IPS_SRC_NAT_DONE is not set, that should also fix it. Could you try this patch please?
net/netfilter/nf_nat_core.c +2 -0
@@ -481,6 +481,8 @@ static int nf_nat_proto_clean(struct nf_conn *i, void *data)if (!nat)return 0 ; +if (!(i->status &IPS_SRC_NAT_DONE )) +return 0 ;if ((clean->l3proto &&nf_ct_l3num (i) != clean->l3proto) || (clean->l4proto &&nf_ct_protonum (i) != clean->l4proto))return 0 ;
Quoted 2 lines · Patrick McHardy
On Fri, 14 Sep 2012, Pablo Neira Ayuso wrote:
[...cut...]
Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Patrick, any other idea?
[...cut...]
Quoted 6 lines
Quoted 3 lines
Quoted 1 line
We can add
nf_nat_iterate_cleanupthat can iterate over the NAT hashtable to replace current usage ofnf_ct_iterate_cleanup.Lets just bail out when
IPS_SRC_NAT_DONEis not set, that should also fix it. Could you try this patch please?
On Fri, 2012-09-14 at 15:15 +0200, Patrick McHardy wrote:
net/netfilter/nf_nat_core.c +0 -0
net/netfilter/nf_nat_core.c static int nf_nat_proto_clean(struct nf_conn *i,
Quoted 4 lines
index 29d4452..8b5d220 100644
net/netfilter/nf_nat_core.c +0
-0@@ -481,6 +481,8 @@ static int nf_nat_proto_clean(struct nf_conn *i,
void *data)
net/netfilter/nf_nat_core.c static int nf_nat_proto_clean(struct nf_conn *i, · line 488
Quoted 9 lines
net/netfilter/nf_nat_core.c +2
-0if (!nat)return 0 ; +if (!(i->status &IPS_SRC_NAT_DONE )) +return 0 ;if ((clean->l3proto &&nf_ct_l3num (i) != clean->l3proto) || (clean->l4proto &&nf_ct_protonum (i) != clean->l4proto))return 0 ;
No it does not work :-(
[ 1216.310146] general protection fault: 0000 [#1] SMP [ 1216.311046] Modules linked in: netconsole ip_vs_lblc ip_vs_lc ip_vs_rr ip_vs libcrc32c ipt_MASQUERADE nf_nat_ipv4(-) nf_nat iptable_mangle xt_mark ip6table_mangle xt_LOG ip6table_filter ip6_tables virtio_balloon virtio_net [last unloaded: iptable_nat] [ 1216.311046] CPU 1 [ 1216.311046] Pid: 4052, comm: modprobe Not tainted 3.6.0-rc3-test-nat-unload-fix+ #32 Red Hat KVM [ 1216.311046] RIP: 0010:[<ffffffffa002c303>] [<ffffffffa002c303>] nf_nat_proto_clean+0x73/0xd0 [nf_nat] [ 1216.311046] RSP: 0018:ffff88007808fe18 EFLAGS: 00010246 [ 1216.311046] RAX: 0000000000000000 RBX: ffff8800728550c0 RCX: ffff8800756288b0 [ 1216.311046] RDX: dead000000200200 RSI: ffff88007808fe88 RDI: ffffffffa002f208 [ 1216.311046] RBP: ffff88007808fe28 R08: ffff88007808e000 R09: 0000000000000000 [ 1216.311046] R10: dead000000200200 R11: dead000000100100 R12: ffffffff81c6dc00 [ 1216.311046] R13: ffff8800787582b8 R14: ffff880078758278 R15: ffff88007808fe88 [ 1216.311046] FS: 00007f515985d700(0000) GS:ffff88007cd00000(0000) knlGS:0000000000000000 [ 1216.311046] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 1216.311046] CR2: 00007f515986a000 CR3: 000000007867a000 CR4: 00000000000006e0 [ 1216.311046] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 1216.311046] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 1216.311046] Process modprobe (pid: 4052, threadinfo ffff88007808e000, task ffff8800756288b0) [ 1216.311046] Stack: [ 1216.311046] ffff88007808fe68 ffffffffa002c290 ffff88007808fe78 ffffffff815614e3 [ 1216.311046] ffffffff00000000 00000aeb00000246 ffff88007808fe68 ffffffff81c6dc00 [ 1216.311046] ffff88007808fe88 ffffffffa00358a0 0000000000000000 000000000040f5b0 [ 1216.311046] Call Trace: [ 1216.311046] [<ffffffffa002c290>] ? nf_nat_net_exit+0x50/0x50 [nf_nat] [ 1216.311046] [<ffffffff815614e3>] nf_ct_iterate_cleanup+0xc3/0x170 [ 1216.311046] [<ffffffffa002c55a>] nf_nat_l3proto_unregister+0x8a/0x100 [nf_nat] [ 1216.311046] [<ffffffff812a0303>] ? compat_prepare_timeout+0x13/0xb0 [ 1216.311046] [<ffffffffa0035848>] nf_nat_l3proto_ipv4_exit+0x10/0x23 [nf_nat_ipv4] [ 1216.311046] [<ffffffff8109f4a5>] sys_delete_module+0x235/0x2b0 [ 1216.311046] [<ffffffff810b8193>] ? __audit_syscall_entry+0x1b3/0x1f0 [ 1216.311046] [<ffffffff810b8776>] ? __audit_syscall_exit+0x3e6/0x410 [ 1216.311046] [<ffffffff816679e2>] system_call_fastpath+0x16/0x1b [ 1216.311046] Code: 75 6e 0f b6 46 01 84 c0 74 05 3a 42 3e 75 61 80 7e 02 00 74 43 48 c7 c7 08 f2 02 a0 e8 37 3b 63 e1 48 8b 03 48 8b 53 08 48 85 c0 <48> 89 02 74 04 48 89 50 08 48 be 00 02 20 00 00 00 ad de 48 c7 [ 1216.311046] RIP [<ffffffffa002c303>] nf_nat_proto_clean+0x73/0xd0 [nf_nat] [ 1216.311046] RSP <ffff88007808fe18>
Signature · 5 lines
Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer
Quoted 31 lines · Jesper Dangaard Brouer
Quoted 2 lines · Patrick McHardy
On Fri, 14 Sep 2012, Pablo Neira Ayuso wrote:
[...cut...]
Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Patrick, any other idea?
[...cut...]
Quoted 6 lines
Quoted 3 lines
Quoted 1 line
We can add
nf_nat_iterate_cleanupthat can iterate over the NAT hashtable to replace current usage ofnf_ct_iterate_cleanup.Lets just bail out when
IPS_SRC_NAT_DONEis not set, that should also fix it. Could you try this patch please?On Fri, 2012-09-14 at 15:15 +0200, Patrick McHardy wrote:
net/netfilter/nf_nat_core.c +0
-0net/netfilter/nf_nat_core.c static int nf_nat_proto_clean(struct nf_conn *i,
Quoted 4 lines
index 29d4452..8b5d220 100644
net/netfilter/nf_nat_core.c +0
-0@@ -481,6 +481,8 @@ static int nf_nat_proto_clean(struct nf_conn *i,void *data)
net/netfilter/nf_nat_core.c static int nf_nat_proto_clean(struct nf_conn *i, · line 488
Quoted 9 lines
net/netfilter/nf_nat_core.c +2
-0if (!nat)return 0 ; +if (!(i->status &IPS_SRC_NAT_DONE )) +return 0 ;if ((clean->l3proto &&nf_ct_l3num (i) != clean->l3proto) || (clean->l4proto &&nf_ct_protonum (i) != clean->l4proto))return 0 ;
Quoted 1 line
No it does not work :-(
Ok I think I understand the problem now, we're invoking the NAT cleanup callback twice with clean->hash = true, once for each direction of the conntrack.
Does this patch fix the problem?
commit 6c46a3bfb2776ca098565daf7e872a3283d14e0d
Author: Patrick McHardy <[email protected]> Date: Thu Sep 20 08:43:02 2012 +0200
netfilter: nf_nat: fix oops when unloading protocol modules
When unloading a protocol module nf_ct_iterate_cleanup() is used to
remove all conntracks using the protocol from the bysource hash and
clean their NAT sections. Since the conntrack isn't actually killed,
the NAT callback is invoked twice, once for each direction, which
causes an oops when trying to delete it from the bysource hash for
the second time.
The same oops can also happen when removing both an L3 and L4 protocol
since the cleanup function doesn't check whether the conntrack has
already been cleaned up.
Pid: 4052, comm: modprobe Not tainted 3.6.0-rc3-test-nat-unload-fix+ #32 Red Hat KVM
RIP: 0010:[<ffffffffa002c303>] [<ffffffffa002c303>] nf_nat_proto_clean+0x73/0xd0 [nf_nat]
RSP: 0018:ffff88007808fe18 EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff8800728550c0 RCX: ffff8800756288b0
RDX: dead000000200200 RSI: ffff88007808fe88 RDI: ffffffffa002f208
RBP: ffff88007808fe28 R08: ffff88007808e000 R09: 0000000000000000
R10: dead000000200200 R11: dead000000100100 R12: ffffffff81c6dc00
R13: ffff8800787582b8 R14: ffff880078758278 R15: ffff88007808fe88
FS: 00007f515985d700(0000) GS:ffff88007cd00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 00007f515986a000 CR3: 000000007867a000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process modprobe (pid: 4052, threadinfo ffff88007808e000, task ffff8800756288b0)
Stack:
ffff88007808fe68 ffffffffa002c290 ffff88007808fe78 ffffffff815614e3
ffffffff00000000 00000aeb00000246 ffff88007808fe68 ffffffff81c6dc00
ffff88007808fe88 ffffffffa00358a0 0000000000000000 000000000040f5b0
Call Trace:
[<ffffffffa002c290>] ? nf_nat_net_exit+0x50/0x50 [nf_nat]
[<ffffffff815614e3>] nf_ct_iterate_cleanup+0xc3/0x170
[<ffffffffa002c55a>] nf_nat_l3proto_unregister+0x8a/0x100 [nf_nat]
[<ffffffff812a0303>] ? compat_prepare_timeout+0x13/0xb0
[<ffffffffa0035848>] nf_nat_l3proto_ipv4_exit+0x10/0x23 [nf_nat_ipv4]
...
To fix this,
- check whether the conntrack has already been cleaned up in
nf_nat_proto_clean
- change nf_ct_iterate_cleanup() to only invoke the callback function
once for each conntrack (IP_CT_DIR_ORIGINAL).
The second change doesn't affect other callers since when conntracks are
actually killed, both directions are removed from the hash immediately
and the callback is already only invoked once. If it is not killed, the
second callback invocation will always return the same decision not to
kill it.
Reported-by: Jesper Dangaard Brouer <[email protected]>
Signed-off-by: Patrick McHardy <[email protected]>net/netfilter/nf_conntrack_core.c +2 -0
@@ -1224,6 +1224,8 @@ get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data),spin_lock_bh (&nf_conntrack_lock);for (; *bucket < net->ct.htable_size; (*bucket)++) {hlist_nulls_for_each_entry (h, n, &net->ct.hash[*bucket], hnnode) { +if (NF_CT_DIRECTION (h) !=IP_CT_DIR_ORIGINAL ) +continue ; ct =nf_ct_tuplehash_to_ctrack (h);if (iter (ct, data)) goto found;
net/netfilter/nf_nat_core.c +2 -0
@@ -481,6 +481,8 @@ static int nf_nat_proto_clean(struct nf_conn *i, void *data)if (!nat)return 0 ; +if (!(i->status &IPS_SRC_NAT_DONE )) +return 0 ;if ((clean->l3proto &&nf_ct_l3num (i) != clean->l3proto) || (clean->l4proto &&nf_ct_protonum (i) != clean->l4proto))return 0 ;
Jesper Dangaard Brouer
Acked-byQuoted 9 lines · Patrick McHardy
Quoted 7 lines · Jesper Dangaard Brouer
Quoted 2 lines · Patrick McHardy
On Fri, 14 Sep 2012, Pablo Neira Ayuso wrote:
[...cut...]
Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Patrick, any other idea?
[...cut...]
[... (hair)cut(?)...]
Quoted 7 lines
Quoted 1 line
No it does not work :-(
Ok I think I understand the problem now, we're invoking the NAT cleanup callback twice with
clean->hash= true, once for each direction of the conntrack.Does this patch fix the problem?
Yes, it fixes the problem :-)
Quoted 18 lines · Jesper Dangaard Brouer
Quoted 9 lines · Patrick McHardy
Quoted 7 lines · Jesper Dangaard Brouer
Quoted 2 lines · Patrick McHardy
On Fri, 14 Sep 2012, Pablo Neira Ayuso wrote:
[...cut...]
Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Patrick, any other idea?
[...cut...]
[... (hair)cut(?)...]
Quoted 6 lines
Quoted 1 line
No it does not work :-(
Ok I think I understand the problem now, we're invoking the NAT cleanup callback twice with
clean->hash= true, once for each direction of the conntrack.
Quoted 5 lines
Quoted 1 line
Does this patch fix the problem?
Yes, it fixes the problem :-)
Great, thanks for testing.
Quoted 106 lines · Patrick McHardy
Quoted 31 lines · Jesper Dangaard Brouer
Quoted 2 lines · Patrick McHardy
On Fri, 14 Sep 2012, Pablo Neira Ayuso wrote:
[...cut...]
Quoted 2 lines
Quoted 2 lines
Quoted 1 line
Patrick, any other idea?
[...cut...]
Quoted 6 lines
Quoted 3 lines
Quoted 1 line
We can add
nf_nat_iterate_cleanupthat can iterate over the NAT hashtable to replace current usage ofnf_ct_iterate_cleanup.Lets just bail out when
IPS_SRC_NAT_DONEis not set, that should also fix it. Could you try this patch please?On Fri, 2012-09-14 at 15:15 +0200, Patrick McHardy wrote:
net/netfilter/nf_nat_core.c +0
-0net/netfilter/nf_nat_core.c static int nf_nat_proto_clean(struct nf_conn *i,
Quoted 4 lines
index 29d4452..8b5d220 100644
net/netfilter/nf_nat_core.c +0
-0@@ -481,6 +481,8 @@ static int nf_nat_proto_clean(struct nf_conn *i,void *data)
net/netfilter/nf_nat_core.c static int nf_nat_proto_clean(struct nf_conn *i, void *data)
Quoted 9 lines
net/netfilter/nf_nat_core.c +2
-0if (!nat)return 0 ; +if (!(i->status &IPS_SRC_NAT_DONE )) +return 0 ;if ((clean->l3proto &&nf_ct_l3num (i) != clean->l3proto) || (clean->l4proto &&nf_ct_protonum (i) != clean->l4proto))return 0 ;Quoted 1 line
No it does not work :-(
Ok I think I understand the problem now, we're invoking the NAT cleanup callback twice with
clean->hash= true, once for each direction of the conntrack.Does this patch fix the problem?
commit
6c46a3bfb2776ca098565daf7e872a3283d14e0dAuthor: Patrick McHardy <[email protected]> Date: Thu Sep 20 08:43:02 2012 +0200netfilter: nf_nat: fix oops when unloading protocol modules When unloading a protocol module nf_ct_iterate_cleanup() is used to remove all conntracks using the protocol from the bysource hash and clean their NAT sections. Since the conntrack isn't actually killed, the NAT callback is invoked twice, once for each direction, which causes an oops when trying to delete it from the bysource hash for the second time. The same oops can also happen when removing both an L3 and L4 protocol since the cleanup function doesn't check whether the conntrack has already been cleaned up. Pid: 4052, comm: modprobe Not tainted 3.6.0-rc3-test-nat-unload-fix+ #32 Red Hat KVM RIP: 0010:[<ffffffffa002c303>] [<ffffffffa002c303>] nf_nat_proto_clean+0x73/0xd0 [nf_nat] RSP: 0018:ffff88007808fe18 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff8800728550c0 RCX: ffff8800756288b0 RDX: dead000000200200 RSI: ffff88007808fe88 RDI: ffffffffa002f208 RBP: ffff88007808fe28 R08: ffff88007808e000 R09: 0000000000000000 R10: dead000000200200 R11: dead000000100100 R12: ffffffff81c6dc00 R13: ffff8800787582b8 R14: ffff880078758278 R15: ffff88007808fe88 FS: 00007f515985d700(0000) GS:ffff88007cd00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 00007f515986a000 CR3: 000000007867a000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process modprobe (pid: 4052, threadinfo ffff88007808e000, task ffff8800756288b0) Stack: ffff88007808fe68 ffffffffa002c290 ffff88007808fe78 ffffffff815614e3 ffffffff00000000 00000aeb00000246 ffff88007808fe68 ffffffff81c6dc00 ffff88007808fe88 ffffffffa00358a0 0000000000000000 000000000040f5b0 Call Trace: [<ffffffffa002c290>] ? nf_nat_net_exit+0x50/0x50 [nf_nat] [<ffffffff815614e3>] nf_ct_iterate_cleanup+0xc3/0x170 [<ffffffffa002c55a>] nf_nat_l3proto_unregister+0x8a/0x100 [nf_nat] [<ffffffff812a0303>] ? compat_prepare_timeout+0x13/0xb0 [<ffffffffa0035848>] nf_nat_l3proto_ipv4_exit+0x10/0x23 [nf_nat_ipv4] ... To fix this, - check whether the conntrack has already been cleaned up in nf_nat_proto_clean - change nf_ct_iterate_cleanup() to only invoke the callback function once for each conntrack (IP_CT_DIR_ORIGINAL). The second change doesn't affect other callers since when conntracks are actually killed, both directions are removed from the hash immediately and the callback is already only invoked once. If it is not killed, the second callback invocation will always return the same decision not to kill it. Reported-by: Jesper Dangaard Brouer <[email protected]> Signed-off-by: Patrick McHardy <[email protected]>net/netfilter/nf_conntrack_core.c +0
-0
net/netfilter/nf_conntrack_core.c get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data), · line 1228
Quoted 5 lines
net/netfilter/nf_conntrack_core.c +2
-0@@ -1224,6 +1224,8 @@ get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data),spin_lock_bh (&nf_conntrack_lock);for (; *bucket < net->ct.htable_size; (*bucket)++) {hlist_nulls_for_each_entry (h, n, &net->ct.hash[*bucket], hnnode) { +if (NF_CT_DIRECTION (h) !=IP_CT_DIR_ORIGINAL ) +continue ;
I think this will make the deletion of entries via `conntrack -F' slowier as we'll have to iterate over more entries (we won't delete entries for the reply tuple).
I think I prefer Florian's patch, it's fairly small and it does not change the current nf_ct_iterate behaviour or adding some nf_nat_iterate cleanup.
Quoted 16 lines
net/netfilter/nf_conntrack_core.c +0
-0ct =nf_ct_tuplehash_to_ctrack (h);if (iter (ct, data)) goto found;net/netfilter/nf_nat_core.c +2
-0@@ -481,6 +481,8 @@ static int nf_nat_proto_clean(struct nf_conn *i, void *data)if (!nat)return 0 ; +if (!(i->status &IPS_SRC_NAT_DONE )) +return 0 ;if ((clean->l3proto &&nf_ct_l3num (i) != clean->l3proto) || (clean->l4proto &&nf_ct_protonum (i) != clean->l4proto))return 0 ;
Quoted 6 lines
Quoted 6 lines
net/netfilter/nf_conntrack_core.c +0
-0@@ -1224,6 +1224,8 @@ get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data),spin_lock_bh (&nf_conntrack_lock);
Quoted 8 lines
net/netfilter/nf_conntrack_core.c get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data), · line 1228
Quoted 4 lines
net/netfilter/nf_conntrack_core.c +2
-0for (; *bucket < net->ct.htable_size; (*bucket)++) {hlist_nulls_for_each_entry (h, n, &net->ct.hash[*bucket], hnnode) { +if (NF_CT_DIRECTION (h) !=IP_CT_DIR_ORIGINAL ) +continue ;I think this will make the deletion of entries via `conntrack -F' slowier as we'll have to iterate over more entries (we won't delete entries for the reply tuple).
Slightly maybe, but I doubt it makes much of a difference.
Quoted 3 lines
I think I prefer Florian's patch, it's fairly small and it does not change the current
nf_ct_iteratebehaviour or adding somenf_nat_iteratecleanup.
I don't think I've received it. Could you forward it to me please?
Quoted 17 lines · Patrick McHardy
Quoted 8 lines
Quoted 8 lines
net/netfilter/nf_conntrack_core.c +0
-0@@ -1224,6 +1224,8 @@ get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data),spin_lock_bh (&nf_conntrack_lock);for (; *bucket < net->ct.htable_size; (*bucket)++) {Quoted 8 lines
net/netfilter/nf_conntrack_core.c get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data), · line 1228
Quoted 4 lines
net/netfilter/nf_conntrack_core.c +2
-0hlist_nulls_for_each_entry (h, n, &net->ct.hash[*bucket], hnnode) { +if (NF_CT_DIRECTION (h) !=IP_CT_DIR_ORIGINAL ) +continue ;I think this will make the deletion of entries via `conntrack -F' slowier as we'll have to iterate over more entries (we won't delete entries for the reply tuple).
Quoted 7 lines
Slightly maybe, but I doubt it makes much of a difference.
Quoted 3 lines
I think I prefer Florian's patch, it's fairly small and it does not change the current
nf_ct_iteratebehaviour or adding somenf_nat_iteratecleanup.I don't think I've received it. Could you forward it to me please?
Florian forwarded the patch to me. While it fixes the problem, it is a workaround and it certainly is inelegant to do the list_del_rcu_init() and memset up to *four* times for a single conntrack.
The correct thing IMO is to invoke the callbacks exactly once per conntrack, either through my nf_ct_iterate_cleanup() change or through a new iteration function for callers that don't kill conntracks. As soon as we start generating events for NAT section cleanup this will be needed in any case.
Unless I'm missing something, conntrack flushing is also a really rare operation anyways and for large tables where this might make a small difference will take a quite large time anyway.
Pablo Neira Ayuso
maintainerQuoted 38 lines · Patrick McHardy
Quoted 18 lines · Patrick McHardy
Quoted 9 lines
Quoted 9 lines
net/netfilter/nf_conntrack_core.c +0
-0@@ -1224,6 +1224,8 @@ get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data),spin_lock_bh (&nf_conntrack_lock);for (; *bucket < net->ct.htable_size; (*bucket)++) {Quoted 8 lines
net/netfilter/nf_conntrack_core.c get_next_corpse(struct net *net, int (*iter)(struct nf_conn *i, void *data), · line 1228
Quoted 4 lines
net/netfilter/nf_conntrack_core.c +2
-0hlist_nulls_for_each_entry (h, n, &net->ct.hash[*bucket], hnnode) { +if (NF_CT_DIRECTION (h) !=IP_CT_DIR_ORIGINAL ) +continue ;I think this will make the deletion of entries via `conntrack -F' slowier as we'll have to iterate over more entries (we won't delete entries for the reply tuple).
Quoted 7 lines
Slightly maybe, but I doubt it makes much of a difference.
Quoted 3 lines
I think I prefer Florian's patch, it's fairly small and it does not change the current
nf_ct_iteratebehaviour or adding somenf_nat_iteratecleanup.I don't think I've received it. Could you forward it to me please?
Florian forwarded the patch to me. While it fixes the problem, it is a workaround and it certainly is inelegant to do the
list_del_rcu_init()and memset up to *four* times for a single conntrack.The correct thing IMO is to invoke the callbacks exactly once per conntrack, either through my
nf_ct_iterate_cleanup()change or through a new iteration function for callers that don't kill conntracks. As soon as we start generating events for NAT section cleanup this will be needed in any case.
Quoted 3 lines
Unless I'm missing something, conntrack flushing is also a really rare operation anyways and for large tables where this might make a small difference will take a quite large time anyway.
Makes sense. And we can revisit this to improve it later.
I'll take this patch. I'll send a batch with updates for the nf-nat thin asap.
Thanks a lot Patrick.
Quoted 2 lines · Pablo Neira Ayuso
Quoted 1 line · Patrick McHardy
On Thu, 20 Sep 2012, Patrick McHardy wrote:
[cut]
(discussion of fixes by Patrick and Florian)
(...settling on Patricks second patch)
Quoted 4 lines
Makes sense. And we can revisit this to improve it later.
I'll take this patch. I'll send a batch with updates for the nf-nat thin asap.
What git tree is that?
I'm trying to work off Pablo's nf-next tree (for my IPVS changes):
git://1984.lsi.us.es/nf-next
But I don't see the patch in that tree ...yet.
Notice, the bug is also present in DaveM's net-next tree. (I know I stated earlier that it didn't affect net-next, but I just forgot to select the new netfilter .config options for nat)
Signature · 43 lines
Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer Oops output, for the sake of completeness: ------------------------------------------ [ 866.878092] general protection fault: 0000 [#1] SMP [ 866.878986] Modules linked in: netconsole ip_vs_lblc ip_vs_lc ip_vs_rr ip_vs libcrc32c ipt_MASQUERADE nf_nat_ipv4(-) nf_nat iptable_mangle xt_mark ip6table_mangle xt_LOG ip6table_filter ip6_tables virtio_net virtio_balloon [last unloaded: iptable_nat] [ 866.879045] CPU 0 [ 866.879045] Pid: 4053, comm: modprobe Not tainted 3.6.0-rc5-net-next-sysctl-tcp+ #13 Red Hat KVM [ 866.879045] RIP: 0010:[<ffffffffa002c2dd>] [<ffffffffa002c2dd>] nf_nat_proto_clean+0x6d/0xc0 [nf_nat] [ 866.879045] RSP: 0018:ffff880078a41e18 EFLAGS: 00010246 [ 866.879045] RAX: 0000000000000000 RBX: ffff880079142500 RCX: dead000000200200 [ 866.879045] RDX: dead000000200200 RSI: ffff880078a41e88 RDI: ffffffffa002f268 [ 866.879045] RBP: ffff880078a41e28 R08: ffff880078a40000 R09: 0000000000000002 [ 866.879045] R10: 0000000000000000 R11: 0000000000000246 R12: ffffffff81c6db40 [ 866.879045] R13: ffff880037d8f008 R14: ffff880037d8f000 R15: ffff880078a41e88 [ 866.879045] FS: 00007fc30c11a700(0000) GS:ffff88007cc00000(0000) knlGS:0000000000000000 [ 866.879045] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 866.879045] CR2: 00007fc30c127000 CR3: 00000000780ef000 CR4: 00000000000006f0 [ 866.879045] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 866.879045] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [ 866.879045] Process modprobe (pid: 4053, threadinfo ffff880078a40000, task ffff88007a379650) [ 866.879045] Stack: [ 866.879045] ffff880078a41e68 ffffffffa002c270 ffff880078a41e78 ffffffff81541413 [ 866.879045] ffffffff00000000 0000147578a40303 ffff880078a41e68 ffffffff81c6db40 [ 866.879045] ffff880078a41e88 ffffffffa00358a0 0000000000000000 000000000040f5b0 [ 866.879045] Call Trace: [ 866.879045] [<ffffffffa002c270>] ? nf_nat_net_exit+0x50/0x50 [nf_nat] [ 866.879045] [<ffffffff81541413>] nf_ct_iterate_cleanup+0xc3/0x170 [ 866.879045] [<ffffffffa002c50a>] nf_nat_l3proto_unregister+0x8a/0x100 [nf_nat] [ 866.879045] [<ffffffff81290303>] ? proc_keys_next+0x23/0x60 [ 866.879045] [<ffffffffa0035848>] nf_nat_l3proto_ipv4_exit+0x10/0x23 [nf_nat_ipv4] [ 866.879045] [<ffffffff810988b5>] sys_delete_module+0x235/0x2b0 [ 866.879045] [<ffffffff810b38c3>] ? __audit_syscall_entry+0x1b3/0x1f0 [ 866.879045] [<ffffffff810b36e6>] ? __audit_syscall_exit+0x3e6/0x410 [ 866.879045] [<ffffffff81640122>] system_call_fastpath+0x16/0x1b [ 866.879045] Code: 75 6c 0f b6 46 01 84 c0 74 05 3a 42 3e 75 5f 80 7e 02 00 74 41 48 c7 c7 68 f2 02 a0 e8 1d c5 60 e1 48 8b 03 48 8b 53 08 48 85 c0 <48> 89 02 74 04 48 89 50 08 48 be 00 02 20 00 00 00 ad de 48 c7 [ 866.879045] RIP [<ffffffffa002c2dd>] nf_nat_proto_clean+0x6d/0xc0 [nf_nat] [ 866.879045] RSP <ffff880078a41e18>
Pablo Neira Ayuso
maintainerQuoted 12 lines · Jesper Dangaard Brouer
Quoted 2 lines · Pablo Neira Ayuso
Quoted 1 line · Patrick McHardy
On Thu, 20 Sep 2012, Patrick McHardy wrote:
[cut]
(discussion of fixes by Patrick and Florian)
(...settling on Patricks second patch)Quoted 4 lines
Makes sense. And we can revisit this to improve it later.
I'll take this patch. I'll send a batch with updates for the nf-nat thin asap.
Quoted 6 lines
What git tree is that?
I'm trying to work off Pablo's nf-next tree (for my IPVS changes):
git://1984.lsi.us.es/nf-nextBut I don't see the patch in that tree ...yet.
I didn't push it yet, will do asap.
Quoted 3 lines
Notice, the bug is also present in DaveM's net-next tree. (I know I stated earlier that it didn't affect net-next, but I just forgot to select the new netfilter .config options for nat)
Pablo Neira Ayuso
maintainerQuoted 13 lines · Pablo Neira Ayuso
Quoted 12 lines · Jesper Dangaard Brouer
Quoted 2 lines · Pablo Neira Ayuso
Quoted 1 line · Patrick McHardy
On Thu, 20 Sep 2012, Patrick McHardy wrote:
[cut]
(discussion of fixes by Patrick and Florian)
(...settling on Patricks second patch)Quoted 4 lines
Makes sense. And we can revisit this to improve it later.
I'll take this patch. I'll send a batch with updates for the nf-nat thin asap.
Quoted 8 lines
Quoted 6 lines
What git tree is that?
I'm trying to work off Pablo's nf-next tree (for my IPVS changes):
git://1984.lsi.us.es/nf-nextBut I don't see the patch in that tree ...yet.
I didn't push it yet, will do asap.
Done.
You may require git pull --rebase to get your patches up on the git head.
On Wed, 2012-09-12 at 23:36 +0200, Florian Westphal wrote:
[...cut...]
Quoted 24 lines
On module removal
nf_nat_ipv4callsnf_iterate_cleanupwhich invokesnf_nat_proto_clean()for each conntrack. That will then callhlist_del_rcu(&nat->bysource)using eachs conntracks nat ext area.Problem is that
nf_nat_proto_clean()is called multiple times for the same conntrack:
a)nf_ct_iterate_cleanup()returns each ct twice (origin, reply)
b) we call it both for l3 and for l4 protocol idsWe barf in
hlist_del_rcuthe 2nd time because ->pprev is poisoned.This was introduced with the ipv6 nat patches.
net/netfilter/nf_nat_core.c +1
-1@@ -487,7 +487,7 @@ static int nf_nat_proto_clean(struct nf_conn *i, void *data)if (clean->hash) {spin_lock_bh (&nf_nat_lock); -hlist_del_rcu (&nat->bysource); +hlist_del_init_rcu (&nat->bysource);spin_unlock_bh (&nf_nat_lock); }else {
Quoted 2 lines
Would probably avoid it. I guess it would be nicer to only call this once for each ct.
Florian's patch fixes the Oops :-)