Re: Oops in 2.6.39 include/net/dst.h: dst_metrics_write_ptr() runningl2tp over ipsec

From: Held Bernhard
Date: Mon Apr 25 2011 - 05:52:25 EST


Am 25.04.2011 10:07, schrieb Eric Dumazet:
From: Held Bernhard<berny156@xxxxxx>

Le lundi 25 avril 2011 Ã 00:37 +0200, Held Bernhard a Ãcrit :
Hi,

I'm starting l2tp over ipsec (racoon, openl2tp) in a little script to
establish a VPN to my company. 2.6.38.x runs fine, but since 2.6.39-rc1
(exactly commit 62fa8a "net: Implement read-only protection and COW'ing
of metrics.") the kernel throws an oops. openl2tp is killed; after a 2nd
start of openl2tp the VPN is established and my PC continues to run
normally. The oops is 100% reproducible.

...

Please tell me if you need more information!

Thanks,
Bernhard
-

Thanks for your report and patch.

Maybe following patch is the way to fix this, please test it.


[PATCH] net: provide cow_metrics() methods to blackhole dst_ops

Since commit 62fa8a846d7d (net: Implement read-only protection and COW'ing
of metrics.) the kernel throws an oops.

[ 101.620985] BUG: unable to handle kernel NULL pointer dereference at
(null)
[ 101.621050] IP: [< (null)>] (null)
[ 101.621084] PGD 6e53c067 PUD 3dd6a067 PMD 0
[ 101.621122] Oops: 0010 [#1] SMP
[ 101.621153] last sysfs file: /sys/devices/virtual/ppp/ppp/uevent
[ 101.621192] CPU 2
[ 101.621206] Modules linked in: l2tp_ppp pppox ppp_generic slhc
l2tp_netlink l2tp_core deflate zlib_deflate twofish_x86_64
twofish_common des_generic cbc ecb sha1_generic hmac af_key
iptable_filter snd_pcm_oss snd_mixer_oss snd_seq snd_seq_device loop
snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec
snd_pcm snd_timer snd i2c_i801 iTCO_wdt psmouse soundcore snd_page_alloc
evdev uhci_hcd ehci_hcd thermal
[ 101.621552]
[ 101.621567] Pid: 5129, comm: openl2tpd Not tainted 2.6.39-rc4-Quad #3
Gigabyte Technology Co., Ltd. G33-DS3R/G33-DS3R
[ 101.621637] RIP: 0010:[<0000000000000000>] [< (null)>] (null)
[ 101.621684] RSP: 0018:ffff88003ddeba60 EFLAGS: 00010202
[ 101.621716] RAX: ffff88003ddb5600 RBX: ffff88003ddb5600 RCX:
0000000000000020
[ 101.621758] RDX: ffffffff81a69a00 RSI: ffffffff81b7ee61 RDI:
ffff88003ddb5600
[ 101.621800] RBP: ffff8800537cd900 R08: 0000000000000000 R09:
ffff88003ddb5600
[ 101.621840] R10: 0000000000000005 R11: 0000000000014b38 R12:
ffff88003ddb5600
[ 101.621881] R13: ffffffff81b7e480 R14: ffffffff81b7e8b8 R15:
ffff88003ddebad8
[ 101.621924] FS: 00007f06e4182700(0000) GS:ffff88007fd00000(0000)
knlGS:0000000000000000
[ 101.621971] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 101.622005] CR2: 0000000000000000 CR3: 0000000045274000 CR4:
00000000000006e0
[ 101.622046] DR0: 0000000000000000 DR1: 0000000000000000 DR2:
0000000000000000
[ 101.622087] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7:
0000000000000400
[ 101.622129] Process openl2tpd (pid: 5129, threadinfo
ffff88003ddea000, task ffff88003de9a280)
[ 101.622177] Stack:
[ 101.622191] ffffffff81447efa ffff88007d3ded80 ffff88003de9a280
ffff88007d3ded80
[ 101.622245] 0000000000000001 ffff88003ddebbb8 ffffffff8148d5a7
0000000000000212
[ 101.622299] ffff88003dcea000 ffff88003dcea188 ffffffff00000001
ffffffff81b7e480
[ 101.622353] Call Trace:
[ 101.622374] [<ffffffff81447efa>] ? ipv4_blackhole_route+0x1ba/0x210
[ 101.622415] [<ffffffff8148d5a7>] ? xfrm_lookup+0x417/0x510
[ 101.622450] [<ffffffff8127672a>] ? extract_buf+0x9a/0x140
[ 101.622485] [<ffffffff8144c6a0>] ? __ip_flush_pending_frames+0x70/0x70
[ 101.622526] [<ffffffff8146fbbf>] ? udp_sendmsg+0x62f/0x810
[ 101.622562] [<ffffffff813f98a6>] ? sock_sendmsg+0x116/0x130
[ 101.622599] [<ffffffff8109df58>] ? find_get_page+0x18/0x90
[ 101.622633] [<ffffffff8109fd6a>] ? filemap_fault+0x12a/0x4b0
[ 101.622668] [<ffffffff813fb5c4>] ? move_addr_to_kernel+0x64/0x90
[ 101.622706] [<ffffffff81405d5a>] ? verify_iovec+0x7a/0xf0
[ 101.622739] [<ffffffff813fc772>] ? sys_sendmsg+0x292/0x420
[ 101.622774] [<ffffffff810b994a>] ? handle_pte_fault+0x8a/0x7c0
[ 101.622810] [<ffffffff810b76fe>] ? __pte_alloc+0xae/0x130
[ 101.622844] [<ffffffff810ba2f8>] ? handle_mm_fault+0x138/0x380
[ 101.622880] [<ffffffff81024af9>] ? do_page_fault+0x189/0x410
[ 101.622915] [<ffffffff813fbe03>] ? sys_getsockname+0xf3/0x110
[ 101.622952] [<ffffffff81450c4d>] ? ip_setsockopt+0x4d/0xa0
[ 101.622986] [<ffffffff813f9932>] ? sockfd_lookup_light+0x22/0x90
[ 101.623024] [<ffffffff814b61fb>] ? system_call_fastpath+0x16/0x1b
[ 101.623060] Code: Bad RIP value.
[ 101.623090] RIP [< (null)>] (null)
[ 101.623125] RSP<ffff88003ddeba60>
[ 101.623146] CR2: 0000000000000000
[ 101.650871] ---[ end trace ca3856a7d8e8dad4 ]---
[ 101.651011] __sk_free: optmem leakage (160 bytes) detected.

The oops happens in dst_metrics_write_ptr()
include/net/dst.h:124: return dst->ops->cow_metrics(dst, p);

dst->ops->cow_metrics is NULL and causes the oops.

Provide cow_metrics() methods, like we did in commit 214f45c91bb
(net: provide default_advmss() methods to blackhole dst_ops)

Signed-off-by: Held Bernhard<berny156@xxxxxx>
Signed-off-by: Eric Dumazet<eric.dumazet@xxxxxxxxx>
---
net/ipv4/route.c | 7 +++++++
net/ipv6/route.c | 7 +++++++
2 files changed, 14 insertions(+)
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index c1acf69..99e6e4b 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -2690,6 +2690,12 @@ static void ipv4_rt_blackhole_update_pmtu(struct dst_entry *dst, u32 mtu)
{
}

+static u32 *ipv4_rt_blackhole_cow_metrics(struct dst_entry *dst,
+ unsigned long old)
+{
+ return NULL;
+}
+
static struct dst_ops ipv4_dst_blackhole_ops = {
.family = AF_INET,
.protocol = cpu_to_be16(ETH_P_IP),
@@ -2698,6 +2704,7 @@ static struct dst_ops ipv4_dst_blackhole_ops = {
.default_mtu = ipv4_blackhole_default_mtu,
.default_advmss = ipv4_default_advmss,
.update_pmtu = ipv4_rt_blackhole_update_pmtu,
+ .cow_metrics = ipv4_rt_blackhole_cow_metrics,
};

struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_orig)
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 843406f..93b10e6 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -153,6 +153,12 @@ static void ip6_rt_blackhole_update_pmtu(struct dst_entry *dst, u32 mtu)
{
}

+static u32 *ip6_rt_blackhole_cow_metrics(struct dst_entry *dst,
+ unsigned long old)
+{
+ return NULL;
+}
+
static struct dst_ops ip6_dst_blackhole_ops = {
.family = AF_INET6,
.protocol = cpu_to_be16(ETH_P_IPV6),
@@ -161,6 +167,7 @@ static struct dst_ops ip6_dst_blackhole_ops = {
.default_mtu = ip6_blackhole_default_mtu,
.default_advmss = ip6_default_advmss,
.update_pmtu = ip6_rt_blackhole_update_pmtu,
+ .cow_metrics = ip6_rt_blackhole_cow_metrics,
};

static const u32 ip6_template_metrics[RTAX_MAX] = {

Your patch works flawlessly.

Thanks for the quick response!

Bernhard
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/