ipmr, ip6mr: Fix several info leaks
Subsystems (4)
Ido Schimmel
maintainercover ipmr, ip6mr: Fix several info leaks
ip{,6}mr_cache_report() copy the IP header of a packet that triggered an upcall to a report packet that is delivered to user space. The header is not always fully initialized or present, resulting in info leaks flagged by KMSAN.
Patch 1 fixes an info leak in the RTM_GETROUTE path, caused by the fact that ipmr_get_route() does not fully initialize the IP header of the packet that it creates. IPv6 does not suffer from this problem.
Patch 2 fixes info leaks triggered by packets injected into pim{,6}reg devices via AF_PACKET sockets. The kernel does not validate that these are IP packets that carry an IP header before calling ip{,6}mr_cache_report().
Both issues were reported by Sashiko and reproduced by me. See KMSAN traces in the commit messages.
As far as I can tell, all other callers call these functions from the IP input / output paths and therefore provide a valid IP packet, with the exception of ip6mr_get_route() which fully initializes the header of the packet it creates.
Ido Schimmel (2):
ipmr: fully initialize the synthetic IP header in ipmr_get_route() ipmr, ip6mr: validate packets injected into the register vif on xmit
net/ipv4/ipmr.c | 7 +++++++ net/ipv6/ip6mr.c | 2 +- 2 files changed, 8 insertions(+), 1 deletion(-)
Signature · 1 line
2.56.0
Ido Schimmel
maintainerPATCH 1/2 ipmr: fully initialize the synthetic IP header in ipmr_get_route()
For an unresolved multicast route query, ipmr_get_route() builds a synthetic IP header in the headroom obtained from skb_realloc_headroom(), which does not zero the memory it allocates. Only a few fields are initialized, leaving others as stale heap contents.
ipmr_cache_report() copies this header into the IGMPMSG_NOCACHE report that is queued to the mroute socket, so a recvmsg() on that socket leaks the uninitialized bytes to user space: the 'tos' field [1] and the 'id' and 'frag_off' fields [2].
Fix by zeroing the IP header before filling in the few fields that are used, so that the entire header is initialized, as in ip6mr_get_route().
[1]
BUG: KMSAN: kernel-infoleak in _copy_to_iter raw_recvmsg
Uninit was stored to memory at:
ipmr_cache_report
ipmr_cache_unresolved
ipmr_get_route
Uninit was created at:
pskb_expand_head
skb_realloc_headroom
ipmr_get_route
Byte 1 of 28 is uninitialized
[2]
BUG: KMSAN: kernel-infoleak in _copy_to_iter raw_recvmsg
Uninit was stored to memory at:
ipmr_cache_report
ipmr_cache_unresolved
ipmr_get_route
Uninit was created at:
pskb_expand_head
skb_realloc_headroom
ipmr_get_route
Bytes 4-7 of 28 are uninitialized
--- net/ipv4/ipmr.c | 1 + 1 file changed, 1 insertion(+)
net/ipv4/ipmr.c +1 -0
@@ -2506,6 +2506,7 @@ int ipmr_get_route(struct net *net, struct sk_buff *skb,skb_push (skb2,sizeof (struct iphdr ));skb_reset_network_header (skb2); iph =ip_hdr (skb2); +memset (iph,0 ,sizeof (*iph)); iph->ihl =sizeof (struct iphdr ) >>2 ; iph->saddr = saddr; iph->daddr = daddr;
Signature · 1 line
2.56.0
Re: [PATCH net 1/2] ipmr: fully initialize the synthetic IP header in ipmr_get_route()
Quoted 62 lines · Ido Schimmel
For an unresolved multicast route query,
ipmr_get_route()builds a synthetic IP header in the headroom obtained fromskb_realloc_headroom(), which does not zero the memory it allocates. Only a few fields are initialized, leaving others as stale heap contents.
ipmr_cache_report()copies this header into theIGMPMSG_NOCACHEreport that is queued to the mroute socket, so arecvmsg()on that socket leaks the uninitialized bytes to user space: the 'tos' field [1] and the 'id' and 'frag_off' fields [2].Fix by zeroing the IP header before filling in the few fields that are used, so that the entire header is initialized, as in
ip6mr_get_route().[1]
BUG: KMSAN: kernel-infoleak in _copy_to_iter raw_recvmsgUninit was stored to memory at:
ipmr_cache_report
ipmr_cache_unresolved
ipmr_get_routeUninit was created at:
pskb_expand_head
skb_realloc_headroom
ipmr_get_routeByte 1 of 28 is uninitialized
[2]
BUG: KMSAN: kernel-infoleak in _copy_to_iter raw_recvmsgUninit was stored to memory at:
ipmr_cache_report
ipmr_cache_unresolved
ipmr_get_routeUninit was created at:
pskb_expand_head
skb_realloc_headroom
ipmr_get_routeBytes 4-7 of 28 are uninitialized
Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")Reported-by Sashiko <[email protected]>Assisted-by LLMSigned-off-by Ido Schimmel <[email protected]>--- net/ipv4/ipmr.c | 1 + 1 file changed, 1 insertion(+)net/ipv4/ipmr.c +0
-0
net/ipv4/ipmr.c int ipmr_get_route(struct net *net, struct sk_buff *skb, · line 2512
Quoted 7 lines
net/ipv4/ipmr.c +1
-0@@ -2506,6 +2506,7 @@ int ipmr_get_route(struct net *net, struct sk_buff *skb,skb_push (skb2,sizeof (struct iphdr ));skb_reset_network_header (skb2); iph =ip_hdr (skb2); +memset (iph,0 ,sizeof (*iph)); iph->ihl =sizeof (struct iphdr ) >>2 ; iph->saddr = saddr; iph->daddr = daddr;
SGTM, I will send this follow up as a separate patch.
net/ipv4/ipmr.c +0 -0
100644
net/ipv4/ipmr.c +0 -0
@@ -550,6 +550,11 @@ static netdev_tx_t reg_vif_xmit(struct sk_buff
*skb, struct net_device *dev)
DEV_STATS_ADD(dev, tx_bytes, skb->len);
DEV_STATS_INC(dev, tx_packets);+ /* skb->cb[] holds qdisc/tc state at this point, not IPCB.
+ * ipmr_cache_report() reads it as IPCB to build IP_PKTINFO.
+ */
+ memset(IPCB(skb), 0, sizeof(*IPCB(skb)));
+
/* Pairs with WRITE_ONCE() in vif_add() and vif_delete() */
ipmr_cache_report(mrt, skb, READ_ONCE(mrt->mroute_reg_vif_num),
IGMPMSG_WHOLEPKT);Ido Schimmel
maintainerPATCH 2/2 ipmr, ip6mr: validate packets injected into the register vif on xmit
In reply to Ido Schimmelreg_vif_xmit() passes the packet to ipmr_cache_report() which copies its IP header into the IGMPMSG_WHOLEPKT report that is queued to the mroute socket. Nothing guarantees that the packet contains a full IP header, so when a short packet is injected on the register device (e.g., using AF_PACKET), the copy reads past the end of the packet and a recvmsg() on the mroute socket leaks the uninitialized bytes to user space [1].
IPv6 has the same problem [2]. Its reg_vif_xmit() does call pskb_inet_may_pull(), but the helper keys off skb->protocol and does not pull anything when the protocol is neither IPv4 nor IPv6.
Fix by dropping packets whose protocol does not match the address family of the register device or that do not contain a full network header.
[1]
BUG: KMSAN: kernel-infoleak in _copy_to_iter raw_recvmsg
Uninit was stored to memory at:
ipmr_cache_report
reg_vif_xmit
packet_sendmsg
Uninit was created at:
__alloc_skb
sock_alloc_send_pskb
packet_sendmsg
Bytes 1-3 of 21 are uninitialized
[2]
BUG: KMSAN: kernel-infoleak in _copy_to_iter rawv6_recvmsg
Uninit was stored to memory at:
ip6mr_cache_report
reg_vif_xmit
packet_sendmsg
Uninit was created at:
__alloc_skb
sock_alloc_send_pskb
packet_sendmsg
Bytes 8-23 of 41 are uninitialized
--- net/ipv4/ipmr.c | 6 ++++++ net/ipv6/ip6mr.c | 2 +- 2 files changed, 7 insertions(+), 1 deletion(-)
net/ipv4/ipmr.c +6 -0
@@ -539,6 +539,12 @@ static netdev_tx_t reg_vif_xmit(struct sk_buff *skb, struct net_device *dev) };int err; +if (skb->protocol !=htons (ETH_P_IP ) || !pskb_inet_may_pull (skb)) { +DEV_STATS_INC (dev, tx_errors); +kfree_skb (skb); +return NETDEV_TX_OK ; + } +rcu_read_lock (); err =ipmr_fib_lookup (net, &fl4, &mrt);if (err <0 ) {
net/ipv6/ip6mr.c +1 -1
@@ -642,7 +642,7 @@ static netdev_tx_t reg_vif_xmit(struct sk_buff *skb, .flowi6_mark = skb->mark, }; -if (!pskb_inet_may_pull (skb)) +if (skb->protocol !=htons (ETH_P_IPV6 ) || !pskb_inet_may_pull (skb)) goto tx_err;rcu_read_lock ();
Signature · 1 line
2.56.0
Re: [PATCH net 2/2] ipmr, ip6mr: validate packets injected into the register vif on xmit
Quoted 8 lines · Ido Schimmel
reg_vif_xmit()passes the packet toipmr_cache_report()which copies its IP header into theIGMPMSG_WHOLEPKTreport that is queued to the mroute socket. Nothing guarantees that the packet contains a full IP header, so when a short packet is injected on the register device (e.g., usingAF_PACKET), the copy reads past the end of the packet and arecvmsg()on the mroute socket leaks the uninitialized bytes to user space [1].
Quoted 7 lines
IPv6 has the same problem [2]. Its
reg_vif_xmit()does callpskb_inet_may_pull(), but the helper keys offskb->protocoland does not pull anything when the protocol is neither IPv4 nor IPv6.Fix by dropping packets whose protocol does not match the address family of the register device or that do not contain a full network header.
Jakub Kicinski
maintainerRe: [PATCH net 2/2] ipmr, ip6mr: validate packets injected into the register vif on xmit
In reply to Ido SchimmelQuoted 1 line · On Sun, 25 Oct 2026 09:17:29 +0200 Ido Schimmel wrote:
Assisted-by LLM
Hi Ido! FWIW these are no longer necessary.
Quoted 22 lines · Ido Schimmel
ip{,6}mr_cache_report() copy the IP header of a packet that triggered an upcall to a report packet that is delivered to user space. The header is not always fully initialized or present, resulting in info leaks flagged by KMSAN.
Patch 1 fixes an info leak in the
RTM_GETROUTEpath, caused by the fact thatipmr_get_route()does not fully initialize the IP header of the packet that it creates. IPv6 does not suffer from this problem.Patch 2 fixes info leaks triggered by packets injected into pim{,6}reg devices via
AF_PACKETsockets. The kernel does not validate that these are IP packets that carry an IP header before calling ip{,6}mr_cache_report().Both issues were reported by Sashiko and reproduced by me. See KMSAN traces in the commit messages.
As far as I can tell, all other callers call these functions from the IP input / output paths and therefore provide a valid IP packet, with the exception of
ip6mr_get_route()which fully initializes the header of the packet it creates.
Quoted 11 lines
Ido Schimmel (2):
ipmr: fully initialize the synthetic IP header inipmr_get_route()ipmr, ip6mr: validate packets injected into the register vif on xmitnet/ipv4/ipmr.c | 7 +++++++ net/ipv6/ip6mr.c | 2 +- 2 files changed, 8 insertions(+), 1 deletion(-)Signature · 1 line
2.56.0
This series was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:
Hello:
This series was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:
Quoted 10 lines · On Sun, 25 Oct 2026 09:17:27 +0200 you wrote:
ip{,6}mr_cache_report() copy the IP header of a packet that triggered an upcall to a report packet that is delivered to user space. The header is not always fully initialized or present, resulting in info leaks flagged by KMSAN.
Patch 1 fixes an info leak in the
RTM_GETROUTEpath, caused by the fact thatipmr_get_route()does not fully initialize the IP header of the packet that it creates. IPv6 does not suffer from this problem.[...]
Here is the summary with links:
- [net,1/2] ipmr: fully initialize the synthetic IP header in ipmr_get_route()
https://git.kernel.org/netdev/net/c/134351edc28e - [net,2/2] ipmr, ip6mr: validate packets injected into the register vif on xmit https://git.kernel.org/netdev/net/c/aa1799e93f5e
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html