Patrick McHardy wrote:
Steve Hill wrote:
This was a configuration mistake on my part and admittedly it
shouldn't work properly - however, it triggered a kernel bug: sending
a packet with the DF flag set which will grow to be > the MTU when
encrypted causes the kernel to generate an ICMP Frag Needed packet,
which got caught by the policy and this triggered the kernel to lock
up hard.
Thanks for tracking this down, we need to unlock the state before
calling icmp_send(). This patch fixes it, it should apply to 2.6.10
if you replace dst_mtu() by dst_pmtu() in the context.
Second try .. this one compiles.
# This is a BitKeeper generated diff -Nru style patch.
#
# ChangeSet
# 2005/03/12 00:59:07+01:00 kaber@xxxxxxxxxxxx
# [XFRM]: Avoid possible deadlock for locally generated ICMP errors
#
# Signed-off-by: Patrick McHardy <kaber@xxxxxxxxx>
#
# net/ipv6/xfrm6_output.c
# 2005/03/12 00:58:59+01:00 kaber@xxxxxxxxxxxx +3 -1
# [XFRM]: Avoid possible deadlock for locally generated ICMP errors
#
# Signed-off-by: Patrick McHardy <kaber@xxxxxxxxx>
#
# net/ipv4/xfrm4_output.c
# 2005/03/12 00:58:59+01:00 kaber@xxxxxxxxxxxx +3 -1
# [XFRM]: Avoid possible deadlock for locally generated ICMP errors
#
# Signed-off-by: Patrick McHardy <kaber@xxxxxxxxx>
#
diff -Nru a/net/ipv4/xfrm4_output.c b/net/ipv4/xfrm4_output.c
--- a/net/ipv4/xfrm4_output.c 2005-03-12 01:03:18 +01:00
+++ b/net/ipv4/xfrm4_output.c 2005-03-12 01:03:18 +01:00
@@ -84,6 +84,7 @@
dst = skb->dst;
mtu = dst_mtu(dst);
if (skb->len > mtu) {
+ spin_unlock_bh(&x->lock);
icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(mtu));
ret = -EMSGSIZE;
}
@@ -111,7 +112,8 @@
if (x->props.mode) {
err = xfrm4_tunnel_check_size(skb);
if (err)
- goto error;
+ /* xfrm4_tunnel_check_size() drops the lock on error */
+ goto error_nolock;
}
xfrm4_encap(skb);
diff -Nru a/net/ipv6/xfrm6_output.c b/net/ipv6/xfrm6_output.c
--- a/net/ipv6/xfrm6_output.c 2005-03-12 01:03:18 +01:00
+++ b/net/ipv6/xfrm6_output.c 2005-03-12 01:03:18 +01:00
@@ -84,6 +84,7 @@
mtu = IPV6_MIN_MTU;
if (skb->len > mtu) {
+ spin_unlock_bh(&x->lock);
icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu, skb->dev);
ret = -EMSGSIZE;
}
@@ -111,7 +112,8 @@
if (x->props.mode) {
err = xfrm6_tunnel_check_size(skb);
if (err)
- goto error;
+ /* xfrm6_tunnel_check_size drops the lock on error */
+ goto error_nolock;
}
xfrm6_encap(skb);
|