net: Fix looping with vrf, xfrms and qdisc on VRF
Subsystems (4)
David Ahern
maintainercover net: Fix looping with vrf, xfrms and qdisc on VRF
From: David Ahern <[email protected]>
Trev reported that use of VRFs with xfrms is looping when a qdisc is added to the VRF device. The combination of xfrm + qdisc is not handled by the VRF driver which lost track that it has already seen the packet.
The XFRM_TRANSFORMED flag is used by the netfilter code for a similar purpose, so re-use for VRF. Patch 1 drops the #ifdef around setting the flag in the xfrm output functions. Patch 2 adds a check to the VRF driver for flag; if set the packet has already passed through the VRF driver once and does not need to recirculated a second time.
This is a day 1 bug with VRFs; stable wise, I would only take this back to 4.14. I have a set of test cases which I will submit to net-next.
David Ahern (2):
xfrm: Always set XFRM_TRANSFORMED in xfrm{4,6}_output_finish
vrf: Check skb for XFRM_TRANSFORMED flag
drivers/net/vrf.c | 6 ++++-- net/ipv4/xfrm4_output.c | 2 -- net/ipv6/xfrm6_output.c | 2 -- 3 files changed, 4 insertions(+), 6 deletions(-)
Signature · 1 line
2.20.1
David Ahern
maintainerPATCH 1/2 xfrm: Always set XFRM_TRANSFORMED in xfrm{4,6}_output_finish
From: David Ahern <[email protected]>
IPSKB_XFRM_TRANSFORMED and IP6SKB_XFRM_TRANSFORMED are skb flags set by xfrm code to tell other skb handlers that the packet has been passed through the xfrm output functions. Simplify the code and just always set them rather than conditionally based on netfilter enabled thus making the flag available for other users.
--- net/ipv4/xfrm4_output.c | 2 -- net/ipv6/xfrm6_output.c | 2 -- 2 files changed, 4 deletions(-)
net/ipv4/xfrm4_output.c +0 -2
@@ -58,9 +58,7 @@ int xfrm4_output_finish(struct sock *sk, struct sk_buff *skb) {memset (IPCB (skb),0 ,sizeof (*IPCB (skb))); -#ifdef CONFIG_NETFILTER IPCB (skb)->flags |=IPSKB_XFRM_TRANSFORMED ; -#endif return xfrm_output (sk, skb); }
net/ipv6/xfrm6_output.c +0 -2
@@ -111,9 +111,7 @@ int xfrm6_output_finish(struct sock *sk, struct sk_buff *skb) {memset (IP6CB (skb),0 ,sizeof (*IP6CB (skb))); -#ifdef CONFIG_NETFILTER IP6CB (skb)->flags |=IP6SKB_XFRM_TRANSFORMED ; -#endif return xfrm_output (sk, skb); }
Signature · 1 line
2.20.1
David Ahern
maintainerPATCH 2/2 vrf: Check skb for XFRM_TRANSFORMED flag
In reply to David AhernFrom: David Ahern <[email protected]>
To avoid a loop with qdiscs and xfrms, check if the skb has already gone through the qdisc attached to the VRF device and then to the xfrm layer. If so, no need for a second redirect.
--- drivers/net/vrf.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/net/vrf.c +4 -2
@@ -474,7 +474,8 @@ static struct sk_buff *vrf_ip6_out(struct net_device *vrf_dev,if (rt6_need_strict (&ipv6_hdr (skb)->daddr))return skb; -if (qdisc_tx_is_default (vrf_dev)) +if (qdisc_tx_is_default (vrf_dev) || +IP6CB (skb)->flags &IP6SKB_XFRM_TRANSFORMED )return vrf_ip6_out_direct (vrf_dev, sk, skb);return vrf_ip6_out_redirect (vrf_dev, skb);
@@ -686,7 +687,8 @@ static struct sk_buff *vrf_ip_out(struct net_device *vrf_dev, ipv4_is_lbcast(ip_hdr(skb)->daddr))return skb; -if (qdisc_tx_is_default (vrf_dev)) +if (qdisc_tx_is_default (vrf_dev) || +IPCB (skb)->flags &IPSKB_XFRM_TRANSFORMED )return vrf_ip_out_direct (vrf_dev, sk, skb);return vrf_ip_out_redirect (vrf_dev, skb);
Signature · 1 line
2.20.1
David Miller
maintainerFrom: David Ahern <[email protected]>
Date: Mon, 20 Apr 2020 17:13:50 -0600
Quoted 13 lines
From: David Ahern <[email protected]>
Trev reported that use of VRFs with xfrms is looping when a qdisc is added to the VRF device. The combination of xfrm + qdisc is not handled by the VRF driver which lost track that it has already seen the packet.
The
XFRM_TRANSFORMEDflag is used by the netfilter code for a similar purpose, so re-use for VRF. Patch 1 drops the #ifdef around setting the flag in the xfrm output functions. Patch 2 adds a check to the VRF driver for flag; if set the packet has already passed through the VRF driver once and does not need to recirculated a second time.
Quoted 3 lines
This is a day 1 bug with VRFs; stable wise, I would only take this back to 4.14. I have a set of test cases which I will submit to net-next.
Series applied and queued up for -stable, thanks David.